After the hotspots are gone: Land use history and grassland plant species diversity in a strongly transformed agricultural landscape

Question: We asked how landscape configuration and present management influence plant species richness and abundance of habitat specialists in grasslands in a ‘modern’(much exploited and transformed) agricultural Swedish landscape. Location: Selaön, south‐eastern Sweden (59°24’ N, 17°10’ E). Methods: Present and past (150 and 50 years ago) landscape pattern was analysed in a 25 km² area. Species richness was investigated in 63 different grassland patches; grazed and abandoned semi‐natural grasslands, and grazed ex‐arable fields. Influence of landscape variables; area, past and present grassland connectivity, present management on total species richness, density and abundance of 25 grassland specialists was analysed. Results: Semi‐natural grasslands (permanent unfertilised pastures or meadows formed by traditional agricultural methods) had declined from 60% 150 years ago to 5% today. There was a significant decline in species richness and density in abandoned semi‐natural grasslands. Total species richness was influenced by present management, size and connectivity to present and past grassland pattern. Landscape variables did not influence species density in grazed semi‐natural grassland suggesting that maintained grazing management makes grassland patches independent of landscape context. The abundance of 16 grassland specialists was mainly influenced by management and to some extent also by landscape variables. Conclusion: Although species richness pattern reflect management and to some extent landscape variables, the response of individual species may be idiosyncratic. The historical signal from past landscapes is weak on present‐day species richness in highly transformed, agricultural landscapes. Generalizations of historical legacies on species diversity in grasslands should consider also highly transformed landscapes and not only landscapes with a high amount of diversity hotspots left.     

Recreating Grasslands in Swedish Rural Landscapes – Effects of Seed Sowing and Management History

Recent loss of plant species richness in Swedish semi-natural grasslands has led to an increase in grassland recreation and restoration. To increase the establishment of declining species favoured by grazing and to re-establish original species richness, seed sowing has been discussed as a conservation tool. In this study, I examined to what extent seed sowing in former arable fields increases species richness and generates a species composition typical of semi-natural grasslands. Six grassland species favoured by grazing (target species) and six generalist species favoured by ceased grazing, were studied in a seed-addition experiment. Four different seed densities were used on four different grassland categories, two grazed former arable fields, one continuously grazed grassland and one abandoned grassland. Target and generalist species emerged in all grassland categories, but seedling emergence was higher in the grazed than in the abandoned grassland. Target species had higher emergence in the two grasslands with the longest grazing continuity. Seedling emergence and frequency of established plants of each target species were positively associated. The largest fraction of seeds germinated at an intermediate sowing density, 20–50 seeds/dm², suggesting that aggregation of seeds positively affects emergence up to a certain threshold. In conclusion, artificial seed sowing may induce the recreation of typical grassland communities on former arable fields, which may be an important contribution to increase the total grassland area and species richness in the landscape.     

Soil arthropod composition differs between old-fields dominated by exotic plant species and remnant native grasslands

Secondary succession after agriculture abandonment (old-fields) is mostly dominated by exotic grass species. Non-native plant invasions may alter soil fauna, potentially inducing plant-soil feedbacks. Despite their importance in nutrient cycling and plant-soil interactions, meso and macrofauna received less attention than bacteria or fungi. Here we compared the composition of the soil arthropod community in native remnants and plant exotic-dominated old-fields grasslands in the Inland Pampa, Argentina. We sampled independent remnants and old-field grassland plots within a 100 km² agricultural landscape to test the hypothesis that the abundance of soil arthropod organisms is related to the quality of the plant biomass, whereas the diversity of the soil biota is related to plant species richness, resulting in a different soil biota composition because of differing plant communities. When compared to non-invaded remnant grasslands, soil activity and soil food-web characteristics of the old-fields sites included: 1. Higher total arthropod abundance, particularly of Isopoda, Pseudoescorpionida and Blattaria; 2. Lower abundance of Hymenoptera and Enthomobryomorpha (Collembola); 3. Lower diversity, and evenness, but similar richness of soil organisms orders; 4. Higher soil respiration rates and soil temperature; and 5. Higher total soil N and K⁺content, but lower soil P content. These results illustrate that soil arthropod composition can vary widely within grasslands patches depending on plant species composition. Also, the more diverse plant community of remnant grasslands supports a more diverse soil biota, although soil activity is slower. Our results support the strong linkage between plant community and soil arthropod composition and suggest that changes in soil biota composition might promote plant-soil feedback interactions inducing the persistence of these alternative grassland states in new agricultural human-modified landscapes.     

Successful restoration of abandoned terraced vineyards and grasslands in Southern Switzerland

Extensively managed semi-natural grasslands represent species-rich habitats and therefore play a key role for the maintenance of biodiversity in agricultural areas. In marginal and poorly accessible areas, the traditional management of grassland is frequently abandoned, which leads to the spread of forest. In Southern Switzerland, terraced vineyards (a special grassland type) and terraced grasslands are part of the cultural heritage and local biodiversity hotspots. Yet, many of them are overgrown by forest. In the past years, several abandoned terraced vineyards and grasslands have been restored by removing the forest, rebuilding the walls and re-introducing the traditional management. We examined restoration success by assessing plant species richness, diversity and species composition in both the aboveground vegetation and soil seed bank in (1) restored, (2) abandoned for 25–50 years, and (3) permanently used areas of six terraced vineyards and six terraced grasslands. Plant species richness and diversity were reduced and species composition altered in the aboveground vegetation of abandoned vineyards and grasslands compared to the permanently used and restored ones. However, species richness, Shannon-diversity and species composition of the aboveground vegetation did not differ between restored and permanently used areas, indicating a successful restoration of the vegetation 10–15 years after restoration. In abandoned vineyards, species richness of plants emerging from the soil seed bank was slightly higher than in permanently used and restored vineyards. No difference in seedling species richness was found between abandoned, permanently used and restored terraced grasslands. Our results showed that the soil seed bank played a minor role for the re-establishment of the above-ground vegetation. We assume that the large species pool in the surroundings and the presence of dispersal vectors are essential for the successful passive restoration of abandoned grassland in this region.     

Density of insect‐pollinated grassland plants decreases with increasing surrounding land‐use intensity

Pollinator declines have raised concerns about the persistence of plant species that depend on insect pollination, in particular by bees, for their reproduction. The impact of pollinator declines remains unknown for species‐rich plant communities found in temperate seminatural grasslands. We investigated effects of land‐use intensity in the surrounding landscape on the distribution of plant traits related to insect pollination in 239 European seminatural grasslands. Increasing arable land use in the surrounding landscape consistently reduced the density of plants depending on bee and insect pollination. Similarly, the relative abundance of bee‐pollination‐dependent plants increased with higher proportions of non‐arable agricultural land (e.g. permanent grassland). This was paralleled by an overall increase in bee abundance and diversity. By isolating the impact of the surrounding landscape from effects of local habitat quality, we show for the first time that grassland plants dependent on insect pollination are particularly susceptible to increasing land‐use intensity in the landscape.     

Effects of land use and landscape patterns on Orthoptera communities in the Western Siberian forest steppe

Across Western Siberia, land use has changed substantially since the collapse of the Soviet Union in 1991: large cropland areas were abandoned and livestock numbers declined. In recent years these trends have partly been reversed, and an intensification of agricultural management has been observed that is still ongoing. We evaluated the impact of land use, as well as effects of landscape patterns and vegetation structure on Orthoptera communities and discuss them as drivers of community composition, species richness and abundance. We sampled Orthoptera using a box-quadrat on ancient grassland, ex-arable grassland (both including different management types: unmanaged, grazed and mown) and cereal fields. Landscape heterogeneity and composition strongly affected species richness and abundance of Orthoptera. Both were higher in grassland than in cropland, but did not differ significantly between ex-arable and ancient grasslands or different management practices. An Indicator Species Analysis revealed differentiation of Orthoptera communities between all management types. On croplands, the number of adult individuals and nymphs was influenced by the proportion of grassland in the surrounding landscape and tillage practices. Conservation tillage is most likely the key factor allowing Orthoptera to reproduce on croplands. After up to 24 years of succession, Orthoptera communities of ex-arable grasslands can be considered as completely recovered, as differences to ancient grasslands were minimal. Besides the continuation of low-intensity management, conservation strategies for this region should consider landscape composition and support habitat heterogeneity like ecotones with hemi-boreal forests in grassland-dominated landscapes.     

Land use history and site location are more important for grassland species richness than local soil properties

Lately there has been a shift in Sweden from grazing species‐rich semi‐natural grasslands towards grazing ex‐arable fields in the modern agricultural landscape. Grazing ex‐arable fields contain a fraction of the plant species richness confined to semi‐natural grasslands. Still, they have been suggested as potential target sites for re‐creation of semi‐natural grasslands. We asked to what extent does fine‐scale variation in soil conditions, management history and site location effect local plant diversity in grazed ex‐arable fields. We examined local soil conditions such as texture, pH, organic carbon, nitrogen (N) and extractable phosphate (P) and effects on plant richness in ten pairs of grazed ex‐fields and neighbouring semi‐natural grasslands in different rural landscapes. Each grassland pair where in the same paddock. A multivariate test showed that site location and land use history explained more of differences in species richness than local soil property variables. Plant species richness was positively associated to grazed ex‐fields with low pH, low N and P levels. Sites with high plant richness in semi‐natural grasslands also had more species in the adjacent grazed ex‐fields, compared to sites neighbouring less species‐rich semi‐natural grasslands. Although both soil properties and species richness were different in grazed ex‐fields compared to semi‐natural grassland, the site location within a landscape, and vicinity to species‐rich grasslands, can override effects of soil properties. In conclusion, if properly located, ex‐arable fields may be an important habitat to maintain plant diversity at larger spatio‐temporal scales and should considered as potential sites for grassland restoration.     

Restoration of Seminatural Grasslands: What is the Impact on Ants?

The number of species‐rich seminatural grasslands in Northern Europe has decreased significantly due to the abandonment of traditional land use practices. To preserve these habitats, an increasing number of abandoned and overgrown grasslands have been restored by cutting down trees and shrubs and reintroducing grazing. These practices are considered a useful tool to recover the species richness of vascular plants, but their impact on other taxa is hardly known. Here we studied ants as one important group of grassland insects. We investigated (1) the effects of restoration of nongrazed and afforested seminatural grasslands, compared to continuously managed reference sites; and (2) the modulating impacts of habitat characteristics and time elapsed since restoration. We found a total of 27 ant species, 11 of these were characteristic of open habitats and seven characteristic of forests. Neither species richness per site nor the number of open‐habitat species, nor the number of forest species differed between restored and reference sites. Yet, within the restored sites, the total species richness and the number of open‐habitat species was positively related to the time since restoration and the percentage of bare rock. High frequencies of most open‐habitat species were associated with low vegetation, older restored sites, and reference sites. Most forest species showed their highest frequencies in tree‐ and shrub‐dominated habitat. We conclude that restoration efforts have been successful in terms of retrieving species richness. A regular and moderate grazing regime subsequent to the restoration is suggested in order to support a high abundance of open‐habitat species.     

Long‐term effects of the abandonment of grazing on steppe‐like grasslands

Question: What are the consequences of grazing abandonment on the Stipa lessingiana dominated steppe‐like grasslands? What is the relative importance of management and environmental factors in causing variation in species composition and abundance in the continuously grazed and abandoned grassland stands? Location: Transylvanian Lowland, Romania. Methods: Repeated vegetation mapping of a grassland stand, where grazing was abandoned 35 years ago; re‐sampling six grassland stands surveyed 29–57 years ago. For revealing long‐term changes in species composition and rank abundance PCoA ordination was applied. The relative importance of management and environmental factors in structuring vegetation were explored by CCA ordination. Diversity, evenness and the relative number and abundance of red‐listed species were compared between managed and abandoned stands. Results: Our results pointed out that grasslands which were formerly grazed and dominated by S. lessingiana, in the long‐term absence of grazing, have been transformed into a S. pulcherrima dominated type. Management, probably by creating bare surfaces and preventing litter accumulation, had the strongest effect on the species composition and abundance in the grasslands. Abandoned grassland stands had lower diversity and evenness compared to continuously grazed stands. While at the same time, the relative number of threatened, rare species did not differ between managed and abandoned sites. Conclusion: Maintaining extensively grazed, as well as un‐managed, Stipa dominated grasslands would be important in order to create various habitat conditions for plant species, especially threatened and rare species, and promote diversity on the landscape scale.     

Effects of creation of open vegetation in abandoned terraced paddy fields on carabid beetle assemblages in temperate Japan

Many paddy fields in the mountainous rural areas of Japan have been abandoned since the 1960s, and forests have regenerated on these sites. In a mountainous area on Sado Island, a large number of abandoned paddies were converted into wetlands and open terrestrial vegetation. In this study, we used pitfall traps to examine the effects of the creation of open vegetation on carabid beetle assemblages by investigating 14 sites spanning five vegetation types: six sites in secondary forests (three coppice forests and three 40‐year‐old regenerating forests on abandoned paddies), three each in clear‐cuts and paddy levees, and two in grasslands. The 14 study sites were clearly separated into two groups different in the species composition of carabid beetles: secondary forest and grassland‐levee groups. The species composition of two clear‐cut sites was similar to that of secondary forests, whereas that of the remaining one clear‐cut site was similar to that of grasslands. Analyses of species responses showed various habitat preferences, e.g., for only coppice forests, for two types of secondary forests, for secondary forests and clear‐cuts, for clear‐cuts and grasslands, and for grasslands or levees, or no clear preference. There were no characteristic species in the regenerating forests. These results suggest that the 40‐year‐old regenerating forests may sustain only a limited subset of the carabid fauna found in coppice forests and that the creation of open vegetation in the abandoned paddies enhances carabid diversity at the landscape level by raising β diversity among the different vegetation types.     

Spontaneous regeneration of dry grasslands on set-aside fields

Dry grasslands are among the most threatened plant communities of Central Europe. We explore the time scale of spontaneous regeneration of dry grasslands on abandoned fields in an area of Central Europe, where also ancient grassland communities occur (Saxony-Anhalt, Germany). On three permanent plots with shallow soils we monitored during 10 years change of species composition and analysed whether spontaneous succession leads to assemblages similar to the ancient dry grassland communities in the direct surroundings. We found that dry grassland species are able to invade the permanent plots and during the 10 years of succession the number of dry grassland species increased. But even after 10 years there was a clear difference between ancient dry grassland communities and the assemblages on the permanent plots. Our findings suggest two important conclusions: First, spontaneous succession on abandoned fields is a cheap possibility for the conservation of some dry grassland species, at least on shallow soils. Second, the time scale of the regeneration process, however, is rather long. Hence, conservation of remnants of ancient grassland communities needs special attention.     

Woody elements benefit bird diversity to a larger extent than semi-natural grasslands in cereal-dominated landscapes

Increasing landscape complexity can mitigate negative effects of agricultural intensification on biodiversity by offering resources complementary to those provided in arable fields. In particular, grazed semi-natural grasslands and woody elements support farmland birds, but little is known about their relative effects on bird diversity and community composition. In addition, the relative importance of local habitat versus landscape composition remains unclear. We investigated how the presence of semi-natural grasslands, the number of woody elements and the composition of the wider agricultural landscape affect bird species richness, true diversity (exponential Shannon diversity) and species composition. Bird communities were surveyed four times on 16 paired transects of 250 m each with 8 transects placed between a crop field and a semi-natural grassland and 8 transects between two crop fields with no semi-natural grasslands in the vicinity. The number of woody elements around transects was selected as an important predictor in all models, having a positive effect on species richness and true diversity, while the local presence of semi-natural grasslands was not selected in the best models. However, species richness and true diversity increased with increasing cover of ley and semi-natural grasslands, whereas species composition was modified by the coverage of winter wheat at the landscape scale. Furthermore, bird species richness, true diversity and species composition differed between sampling dates. As bird diversity benefited from woody elements, rather than from the local presence of semi-natural grasslands as such, it is important to maintain woody structures in farmland. However, the positive effect of grassland at the landscape scale highlights the importance of habitat variability at multiple scales. Because species richness and true diversity were affected by different landscape components compared to species composition, a mosaic of land-use types is needed to achieve multiple conservation goals across agricultural landscapes.     

The Influence of Landscape Grain Size on Butterfly Diversity in Grasslands

The relationship between butterfly diversity and both habitat and landscape variables was studied in two areas of southern Sweden. The habitat quality of the grasslands was similar in the two study areas but the landscape pattern differed in grain size and amount of grassland and forest. Using a transect survey method, a total of 3341 butterflies were observed and 30 taxa identified. We found that both habitat and landscape variables influenced the butterfly diversity of the investigated grasslands. Species composition differed markedly between the two study areas. A study area with a fine-grained landscape pattern, a high cover of semi-natural grassland and many forest edges had twice as many butterfly species but half the number of individuals compared with a coarser-grained study area with larger grasslands widely spread in a matrix of arable fields. The results of our study indicate that both habitat quality and landscape pattern have to be considered when developing conservation strategies for grassland butterflies.     

The early response of subtropical tussock grasslands to restoration treatments

Once widespread, Australia's bluegrass tussock grasslands (dominated by Dichanthium sp.) of the Queensland Central Highlands are now severely endangered. Despite being biodiversity rich and highly valued as low input, nutrient‐dense grazing systems, bluegrass tussock grasslands have suffered extensive clearing and degradation over the last 150 years. Natural recovery of these grasslands is possible but rates of recovery are slow. As such, there is an urgent need to assess practical management strategies to accelerate recovery of these grasslands, with a particular focus on early‐successional stages, when aggressive exotic species are most prevalent. To date, no studies have tested whether commonly used grassland restoration strategies can enhance early‐successional stages and accelerate regeneration in this system. Here, we examine the early short‐term impacts (first two seasons) of two common grassland restoration approaches, with two variations each: direct seeding (single species and low seed diversity) and vegetation clearing (prescribed burning and glyphosate application) across two common starting points: a formerly cropped old field and a historically overgrazed natural grassland. No treatment increased native diversity (Shannon's or richness) in plots but the composition of burned plots in the old field did become more similar to healthy reference sites after two seasons. Burning combined with direct seeding also increased the abundance of the dominant grass, Dichanthium sericeum, toward healthy reference levels within the first two seasons post seeding. This study provides a practical assessment of the short‐term impacts and capacity of common grassland restoration treatments to enhance the recovery of Australia's tussock grassland systems.     

Post-Soviet recovery of grassland vegetation on abandoned fields in the forest steppe zone of Western Siberia

Following the collapse of the Soviet Union in 1991 around 45 million hectares of arable land became abandoned across Russia. Our study focused on the recovery potential and conservation value of grassland vegetation on ex-arable land in the Tyumen region of the Western Siberian grain belt. We compared ex-arable grasslands of different successional stages with ancient grasslands as reference for the final stage of succession along a climatic gradient from the pre-taiga to the forest steppe zone. Plant community composition and species richness of ex-arable land clearly developed towards reference sites over time, but even after 24 years of abandonment, the grassland vegetation had not totally recovered. The γ-diversity of vascular plants was slightly higher on ex-arable land than in ancient grasslands but the mean α-diversity was still moderately lower. A significant proportion of the vegetation of ex-arable land still consisted of ruderal and mesic grassland species and the number and cover of meadow-steppe species was significantly lower than in ancient grasslands. Grazing and time since abandonment positively affected the reestablishment of target grassland species, whereas it was negatively affected by the cover of grasses. In contrast to ex-arable land, the conservation value of arable land is only modest. Therefore, future intensification of land use is most likely less harmful if directed to existing arable land. Re-cultivation of ex-arable land and grassland improvement operations such as seeding of competitive grass species are major threats for the biodiversity of secondary grasslands on ex-arable land in the forest steppe zone of Western Siberia.     

The role of linear mown firebreaks in conserving butterfly diversity: Effects of adjacent vegetation and management

Semi‐natural grasslands are vital for maintaining grassland butterflies in Japan, as well as in Europe. However, severe decline in these grassland environments has recently attracted attention to linear grasslands, such as firebreaks and power‐line corridors, as alternative habitats for grassland insects. We surveyed butterflies in an abandoned grassland and nearby linear mown firebreaks adjacent to different vegetation at the northern foot of Mt. Fuji, central Japan, over 5 successive years, particularly focusing on species on the 2012 Japanese Red List of Threatened Species (“red‐list” species). We found that the firebreaks were consistently higher in species richness and abundance of butterflies than the long‐term abandoned grassland, and that species composition differed among the firebreaks depending on conditions of the adjacent vegetation. The firebreaks surrounded by forests were mainly utilized by forest and edge species, whereas the firebreaks adjacent to the grassland were essential for conserving red‐list grassland species. Thus, only the mown firebreaks adjacent to the grassland were regarded as a high‐quality alternative habitat for many grassland butterflies, but the area was limited. Therefore, creating heterogeneity in the abandoned grassland by infrequent mowing could help conserve grassland butterflies, including red‐list species.     

Near‐natural methods promote restoration of species‐rich grassland vegetation—revisiting a road verge trial after 9 years

The present loss of species‐rich grasslands makes it vital to restore these valuable habitat types, including novel habitat variants such as road verges. Due to the lack of knowledge on long‐term outcomes of restoration initiatives, well‐designed studies comparing different restoration methods are needed. In this study, we examined fine‐scale vegetation recovery patterns over 9 years in a field experiment with several near‐natural restoration methods (adding local seed mixtures, transferring hay from local grasslands using hard or light raking, and natural regeneration) in a road verge. We compared this to standard revegetation (hydroseeding species‐poor commercial seed mixtures). We found major temporal changes in vegetation restored by local seed or hay transfer, before it gradually became more similar to the donor grasslands and seed mixtures, which served as references for the experiment. Natural (spontaneous) regeneration with seed dispersal from surroundings gave similar results, whereas areas revegetated using standard methods became more dissimilar to the reference sites during the study period. The main variation in species composition reflected the contrast between local donor grasslands and seed mixtures and the species‐poor early successional grasslands. We conclude that near‐natural methods (hay transfer and seeding) successfully restored species‐rich grassland, including road verges. This study underlines the importance of comparing several treatments over a sufficiently long period to assess their success in restoring species‐rich grassland.     

Plant species richness in calcareous grasslands as affected by dispersability in space and time

Abstract. Species richness in calcareous grassland is discussed against the background of historical dispersal processes associated with traditional land-use management such as grazing, but also the artificial establishment by hayseed. Important vectors in the traditionally man-made landscape were sheep and cattle or other livestock such as goats. Calcareous grasslands were not only connected to each other but also to other habitats such as villages, forests, arable fields and heathlands by these vectors which could cover large distances (e.g. transhumance shepherding), which is not the case in the current man-made landscape. Species richness after restoration management of abandoned and afforested calcareous grasslands was predicted by using characters of dispersability in space and time. These were the persistence of the species in the vegetation and the diaspore bank after abandonment or afforestation and the dispersal capacity through wind and sheep. The results reveal that reintroduction of sheep grazing is necessary to reestablish the original species richness. The first validation of the prediction of the succession on clear-cut sites and a comparison with data of species composition in abandoned quarries and the surroundings made it obvious that a species' own dispersal capacity in space is very low except for some well wind-dispersed species. Therefore, it is recommended to include and to simulate dispersal processes in future management to be able to restore the original species richness.     

Recovery of plains rough fescue grasslands on reclaimed well sites

Plains rough fescue (Festuca hallii), once dominant in grasslands of the Northern Great Plains, has been reduced to remnants mainly through agricultural and energy sector development. This study assessed the impacts of oil and gas well site disturbances on plains rough fescue grassland to predict successional trends following disturbance. We examined trends in vegetation cover, richness, diversity, and community composition for two construction techniques (topsoil stripping, minimum disturbance), three revegetation methods (agronomic seed mix, native seed mix, natural recovery), and two reclamation scenarios (reclaimed within < 10 yrs; reclaimed within > 10 yrs) relative to adjacent undisturbed prairie (reference sites) over 28 years in 33 grassland sites. Reclamation success was more closely related to methods of construction and revegetation than years since reclamation. Species richness, diversity, both native and non-native species cover, and species composition were similar between undisturbed prairie and areas subject to minimum disturbance and natural recovery. In contrast, undisturbed prairie differed from areas with topsoil stripping and seeding to either agronomic or native species. Plant community composition on minimum disturbance sites with natural recovery was returning to a predisturbed plains rough fescue community within 10 years after reclamation. Impacts of construction method that involved intensive soil handling and seeding with native or non-native seed mixes were disruptive to recovery of fescue grassland. We therefore recommend retaining grassland sod intact through minimum disturbance and utilizing natural recovery as the best option for successful reclamation of native rough fescue grassland after well site disturbance.     

Remnant habitats for grassland species in an abandoned Swedish agricultural landscape

Questions: Which factors influence the persistence of vascular grassland plants in long‐abandoned (at least 50 yr) arable fields and meadows? What might be the implications of current levels of species richness on abandoned arable fields and meadows for future restoration? Location: Forested highlands of Kilsbergen, south central Sweden. Methods: The abundance of all vascular plant species was investigated in three habitat types: former arable fields, hay meadows and outlands (pastures) at 27 farms, abandoned for either approximately 50 yr or 90 yr. Time since abandonment, tree cover, soil depth, degree of soil podsol development, size of the infield area and two measures of connectivity were used as predictors for species richness and species composition. Results: Former outland had denser tree cover, fewer species and fewer grassland species than former arable fields and hay meadows, irrespective of time since abandonment. Former hay meadows and arable fields with a longer time since abandonment were less rich in species, more wooded and had greater podsolization than meadows and fields abandoned at a later stage. Species richness was higher in hay meadows and arable fields at farms with larger infield area and deeper soils compared with farms with smaller infield area and shallower soils. The greatest richness of species and most open habitat were former arable fields at larger farms abandoned 50 yr before the study. Former arable fields had the highest number of grassland species. Conclusion: After 50 yr of abandonment, former arable fields were the most important remnant habitats for grassland species and may be a more promising target for restoration than formerly managed grasslands.