Habitat fragmentation has far-reaching negative impacts on the environment, resulting in biodiversity loss, soil quality degradation and alteration of water availability. In addition, fragmentation can disrupt ecological processes, potentially facilitating the establishment and spread of invasive plants, which can further harm native arthropod communities and alter their ecosystem dynamics. However, the exact nature of these impacts may vary depending on local conditions. We investigated the impact of fragmentation and milkweed invasion on invertebrate communities in sandy grasslands of forest-steppe habitats in Hungary. We selected 30 grasslands in forest-steppe fragments, varying in size (0.2 to 8.7 ha) and connectivity (Hanski’s connectivity index: 0 to 705). We sampled ground-dwelling arthropods, mainly herbivores (true bugs) and predators (spiders), with pitfall traps and pollinators (wild bees) by direct observations along transects in invaded vs. non-invaded patches (min. of 500 m²) of each fragment. We considered arthropod species’ body size (all groups), dispersal ability and feeding (herbivores and predators) and nesting location and social habit (wild bee) traits in our analyses. In non-invaded patches, the number of monophagous herbivores showed an increasing trend, whereas in invaded fragments, there were more polyphagous individuals with increasing connectivity and fragment size. The dispersal ability of predators was lower as connectivity increased in non-invaded patches but higher in patches invaded by milkweed. We found more ground nesting bees in the invaded patches of small fragments than in large fragments, however, we did not find a significant effect in non-invaded patches. In summary, we often found interacting effects of the studied variables, fragmentation and invasions, generally modifying each other’s effect by filtering for opposite trait levels. The primary objective of restoration projects should be restoring habitat of appropriate size and connectivity and eradicating invasive species while concurrently supporting the revival of native species and their ecological relationships. It is essential to employ adaptive management techniques, including continuous monitoring, to effectively tackle the interaction between fragmentation, invasion, and the preservation of biodiversity.

Biodiversity loss, connectivity, functional diversity, habitat degradation, landscape structure, sandy grassland

Land-use changes, such as afforestation of grasslands, urbanisation and agricultural expansion, result in the fragmentation of natural habitats (Fischer and Lindenmayer 2007). Habitat fragmentation per se is independent of habitat loss, for a given amount of habitat, a more fragmented landscape has more, but smaller habitat patches (Fahrig 2017). Many small habitat patches seem to host more species than a few large patches of the same total area (Fahrig 2020). However, loss of habitat amount can negatively impact biodiversity, as it can limit the movement and gene flow of many species, increasing the risk of extinction and altering the species composition within habitat patches (Fahrig 2003; Tremlová and Münzbergová 2007). Conservation efforts may aim to restore habitat amount and connectivity between fragments to mitigate negative impacts (Wilson et al. 2016). Ensuring adequate habitat size and connectivity is crucial for maintaining biodiversity, promoting ecosystem health and supporting the resilience of species and ecosystems in the face of changing environmental conditions (Correa-Ayram et al. 2016).

Fragmentation can pave the way for non-native plant species to establish and spread, which may lead to adverse ecological impacts (Keller et al. 2008; Gutiérrez et al. 2014). Invasive plant species typically compete with native plants for resources like water, nutrients, light, and space or even pollinators, leading to reduced native vegetation diversity and a shift in the structure and composition of plant communities (Crooks 2002). These changes can have cascading effects on the whole ecosystem (Diez et al. 2010), including changes to the food web and the availability of animal nesting and foraging sites. For instance, invasive plants may not provide pollinators with the same quality or quantity of food resources as native plants (Bjerknes et al. 2007). This can lead to a reduction in pollinator densities, which can affect the reproductive success of native plants and the availability of fruits and seeds for their consumers.

Common milkweed (Asclepias syriaca) is native to North America. It is a herbaceous perennial plant that grows 30–180 cm tall and produces clusters of fragrant, pink or purple flowers in the summer. Common milkweed provides food and habitat for a wide variety of wildlife, including many specialists (Spaeth et al. 2022). Therefore, common milkweed has high nature conservation value in its native range; however, its invasiveness should be carefully monitored and managed outside of its native range (Zalai et al. 2017). In Europe, the common milkweed is considered an invasive species in several countries, and it is included in the list of invasive alien species of the European Union (EU list 2017). It was originally introduced to Europe in the 17^th century as an ornamental plant and by beekeepers as a key food resource for honeybees due to its large quantities of nectar and pollen (Bukovinszky et al. 2014). Common milkweed was introduced to Hungary in the 18^th century by beekeepers to increase food availability for their bees and improve the health and productivity of their hives. Milkweed became invasive and spread rapidly in the lowland areas of Central Europe (Bakacsy and Bagi 2020).

Milkweed has a negative effect on the native, habitat specialist plants. The most vulnerable areas to milkweed invasion are those where the sandy soils have been degraded, and habitat disturbance is already threatening the native vegetation of grasslands (e.g. disturbed by overgrazing, Bakacsy and Bagi 2020) and forest plantations (e.g. disturbed by intensive forestry, Ingle et al. 2019). Due to its aggressive growth habit, common milkweed can dominate and displace other plant species. Milkweed has a positive effect on habitat generalist and forest plants (Gallé et al. 2023). Furthermore, it negatively affects grassland species, especially those with low competitive ability (Kelemen et al. 2016; Berki et al. 2023). Therefore, milkweed invasion has a general negative effect on the nature conservation value of invaded habitats. The presence of milkweed also alters vegetation structure, it increases the total cover and vegetation height and decreases bare ground cover (Gallé et al. 2023).

One of the most threatened habitat types by milkweed invasion is the forest steppes in southern Hungary (Bakacsy and Bagi 2020). Forest steppes are a mosaic of forest patches on grasslands (Fig. 1a). They form a distinct vegetation belt, a transition between closed forests and mostly treeless steppes in Eastern Europe and Asia (Bátori et al. 2018). Only small fragments of forest steppe remained in Southern Hungary, where fragment size and habitat connectivity are important drivers of arthropod diversity, especially in the forest-steppe grasslands (Gallé et al. 2022a), generally exerting a stronger effect on habitat specialist than generalist arthropods (Gallé et al. 2023).

Figure 1.

Map of study area showing the habitat fragments and the schematic figure of within-fragment sampling design.

Arthropods are essential parts of ecosystems, fulfilling multiple roles, such as herbivores, pollinators, predators, decomposers and prey for other organisms (Schowalter 2022). Several factors affect arthropods’ species and trait state composition. Focusing on functional traits can identify mechanisms that determine the impact of biodiversity on ecosystem processes (Spasojevic et al. 2018; Gallé and Batáry 2019). Identifying the drivers behind the trait state composition of arthropod communities can help us better understand their role in ecosystems and preserve their biodiversity (Haddad et al. 2008). The proliferation of an invasive plant, such as milkweed, may alter ecosystems, negatively impacting ground-dwelling arthropod populations (Gallé et al. 2015; Ingle et al. 2019). For example, Šeat et al. (2024) found that larger true bugs and species with generalist diets were associated with invasive plants in saline grasslands. Plant invasion may support web builders, reflecting vegetation composition and structural changes (Gomes et al. 2018). Connectivity and larger fragment size support large arthropod species, which can be expected to be good dispersers (Kormann et al. 2015, Korányi et al. 2023). Milkweed may affect wild bees positively, neutrally, and negatively (Szigeti et al. 2020; Kovács-Hostyánszki et al. 2022). The impact of common milkweed on wild bees depends on their traits, the local ecosystem and the abundance and distribution of other food and habitat resources (Gustafson et al. 2023).

In this study, we aimed to reveal the combined effects of milkweed invasion and habitat fragmentation on the functional diversity of three arthropod groups: herbivorous true bugs, wild bees and predatory spiders. We expected that (1) trait state and species composition differ between arthropod assemblages in invaded and non-invaded areas. We expected that varying connectivity and fragment size also shape trait composition: (2) Species dispersing over longer distances are able to populate less connected small fragments, whereas those with poor dispersal ability may be typical of well-connected large fragments; (3) Invasion homogenises food resources; thus species with a generalist diet are expected in invaded patches whereas feeding specialists are expected in areas that are close to natural.

We collected 2797 true bug individuals belonging to 58 species, 854 wild bees belonging to 30 species and 2544 spiders representing 65 species (Suppl. material 1). The CWM of true-bugs body size increased with increasing connectivity in invaded patches but decreased in non-invaded ones. (Table 1, Fig. 2A). Connectivity had a significant positive effect on the wing length of true bugs, as there were more individuals with developed wings in well-connected fragments than in poorly connected fragments (Table 1, Fig. 2B). We found a similar pattern for fragment size and connectivity effect on herbivore diet as it was influenced by both fragment size and connectivity in interaction with invasion. The diet of herbivorous true bugs became more specialised (i.e., the community shifted from polyphagous toward monophagous individuals) in non-invaded patches and increased in invaded fragments with increasing connectivity and fragment size (Table 1, Fig. 2C, D).

Figure 2.

Effects of invasion and landscape variables on arthropod functional trait CWMs (community weighted mean). These are abundance-weighted averages of indices ranging from 0-1 (see Methods) A interacting effect of milkweed invasion and connectivity on herbivore size CWM B connectivity affect dispersal CWM of herbivores C invasion and fragment size effect on herbivore diet CWM D the interacting effect of invasion and on the diet CWMs of herbivores E interacting effect of invasion and fragment size on wild bee species nest height CWM F interacting effect of invasion and fragment size on spider dispersal CWM. Ranged values of connectivity and log-transformed fragment size are plotted. Grey dots show invaded sites and green dots show non-invaded sites. We show fitted regression lines with 95% confidence intervals.

Bee individuals with higher nesting locations were more common in non-invaded patches. We also detected an interaction effect of invasion and fragment size on the CWMs of wild bees’ nesting strategy. We found more ground-nesting wild bees in non-invaded patches in small than large fragments (Table 1, Fig. 3E).

Figure 3.

Non-metric multidimensional scaling ordination (NMDS) of A wild bee (stress = 0.073), and B predator (stress = 0.260) community composition. Sites are indicated with open circles. Red arrow indicates the significant continuous variable, and convex hulls indicate the habitat (C: control (non-invaded), I: invaded). Indicator species of invaded areas are indicated with black dots, and indicator species of non-invaded areas are green dots. Bter: Bombus terrestris (Apidae), Pnig: Pellenes nigrociliatus (Salticidae), Zexi: Zelotes exigus (Gnaphosidae), Hboh: Haplodrassus bohemicus (Gnaphosidae) and Zseg: Zelotes segrex (Gnaphosidae).

The ballooning propensity, the proxy for the dispersal ability of spiders decreased (i.e. the number of poorly dispersing individuals increased) in non-invaded patches as connectivity increased. In contrast, in patches invaded by milkweed, CWM of dispersal ability values increased with increasing connectivity (Table 1, Fig. 3F).

We found significant differences in the wild bee and predator community composition of the invaded and the non-invaded patches (Table 2). We detected a single indicator wild bee species, namely Bombus terrestris, in the invaded habitats (Fig. 3A). In the case of predators, we found one indicator species, the salticid Pellenes nigrociliatus, for the non-invaded and three gnaphosid species, Haplodrassus bohemicus, Zelotes exigus, Zelotes segrex for the invaded fragments (Fig. 3B). No differences in community composition were observed for herbivores (Table 2).

We studied the effect of invasive milkweed, fragment size, and connectivity on different arthropod communities inhabiting a threatened habitat type, the forest steppe. We found an interacting effect of the presence of the invasive species, connectivity and fragment size affected on most arthropod traits. Specifically, our hypothesis (1) was supported: invasion or its interaction with fragment size or connectivity affected the traits of all arthropods studied. Hypothesis (2) was also supported: the presence of the invasive plant shaped the community composition of wild bees and spiders. Hypothesis (3) was supported for true bugs in well-connected fragments: we found more polyphagous herbivores in invaded sites than in non-invaded sites.

Corridors or interconnected patches facilitate large individuals’ movement, increasing the likelihood of their presence, but this effect was not attributed to a single or few species. Large herbivores typically require ample food resources to sustain their larger body size (Tscharntke et al. 2002). Herbivore body size increased with higher connectivity in non-invaded but decreased in invaded patches. This suggests that smaller herbivorous insects might be more common in well-connected habitats. But the presence of invasive plants filters for smaller species and disadvantaging the larger species in poorly connected habitats.The collected true bug species mainly rely on native plants, generally avoiding milkweed, as this plant is poisonous to many herbivores. Only a few European true bug species feed on the milkweed, from the genus Lygaeus. Indeed, we collected more Lygaeus simulans individuals in invaded (34 individuals, 77%) than in non-invaded areas (10 individuals, 22%). Native plants of the forest-steppe grasslands can be more accessible and readily available in non-invaded patches, promoting the occurrence of larger true bug species. If the preferred host plants of a true bug species are negatively affected by invasive plants, this can lead to a decline in the abundance of that true bug species (Crooks 2002). This may occur despite increasing habitat connectivity.

In line with our expectations, large and well-connected fragments supported feeding specialists if fragments were non-invaded (e.g. Fischer and Lindenmayer 2007). The herbivore community shifted towards species with specialist diets in non-invaded patches, but in invaded patches, the generalist species were more frequent with increasing connectivity and fragment size. This suggests that invasive plants can affect herbivores’ feeding specialisation, resulting in more generalists in large and connected fragments. The presence of invasive milkweed can have a negative impact on herbivores both directly and indirectly (Gallé et al. 2023; Korányi et al. 2023). Herbivores encounter invasive plants with chemical compositions that differ significantly from those of native plants (Xiao et al. 2020). Native herbivores are not adapted, or are poorly adapted, to feed on invasive plants. The indirect effect of milkweed on feeding specialisation might have been established through the effect of the invasive plant on native vegetation. Invasive plants negatively affect the native plant communities (Vilà et al. 2011). Consequently, they may reduce the access of specialist herbivores to their native host plants. Milkweed affects the species composition of native vegetation by supporting more generalist plant species, which provide a more diverse diet for generalist true bugs (Gallé et al. 2023). Moreover, milkweed can provide shade through its canopy and litter, potentially alleviating unfavourable abiotic (microclimate) conditions for the germination of specialist plant species (Szitár et al. 2018). Furthermore, the large stems of milkweed change the habitat structure of sandy grasslands, presumably reducing the availability of suitable oviposition sites and compromising herbivores’ survival and reproductive success (Tallamy and Shropshire 2009). The generalist plants in invaded patches might formed a suitable habitat for generalist herbivores; therefore, connectivity and fragment size supported their high abundance.

The availability of suitable nesting sites is crucial for solitary wild bees, making them highly vulnerable to unfavourable environmental factors and the impacts of human-induced changes (Harmon-Threatt 2020). Milkweed was the highest herbaceous plant species in our study, producing large quantities of relatively high stems. Above-ground nesting bees build their nests in holes in plant stems or dead wood (Bihaly et al. 2021). The increased ratio of wild bees nesting on higher vegetation in non-invaded patches means that large milkweed stems did not offer more suitable nesting sites for above-ground nesters. Additionally, the large fragments support a larger population of vegetation-nesters, presumably by offering more nesting sites and more available food resources. Minimising the distance wild bees have to cover between the food resource and nesting location during foraging is essential to allocating energy for their reproduction. Non-invaded patches bear lower average vegetation heights and more bare soil surfaces, which create ideal conditions for ground-nesting species (Harmon-Threatt 2020). The majority of recorded bee species nest in the soil and spend much of their life cycle underground. Their preferences may differ concerning soil type, texture, compaction, moisture, and temperature (Antoine and Forrest 2021). Lower soil moisture and lower ground surface temperatures caused by milkweed invasion (Gallé et al. 2023) could potentially limit ground-nesting species (Pellaton et al. 2024), however ground-nesting bees generally dominated the communities of forest-steppe fragments. Large and well-connected, flower-rich forest-steppe fragments are beneficial for ground-nesting bees (Török et al. 2022). Further, the impact of milkweed invasion on different floral visitor groups may vary, as their diversity is probably influenced by the presence and variety of native flowering resources (Kovács-Hostyánszki et al. 2022). However, it did not affect the size and social behaviour of wild bees.

Well-connected fragments are often viewed as higher-quality habitats with more specialist species. They support a high species richness and abundance of arthropods (Hanski 1998), which are potential food items for predators, provide suitable conditions and high-quality habitats for many spiders including poorly dispersing species (Gallé et al. 2022a). In sandy grasslands, vegetation structure is also among the most important parameters that affects spiders (Carvalho et al. 2011). Specifically, vegetation with complex structures supports many spider species (Gallé et al. 2010). Milkweed increases vegetation complexity, and, therefore might have an indirect positive effect on spiders and change trait state composition of the community. We found that predator dispersal ability decreased with connectivity in non-invaded patches, but increased in patches invaded by milkweed.

Spiders with good dispersal ability can move between fragments more efficiently in landscapes with well-connected fragments, facilitating gene flow and maintaining stable populations (Gallé et al. 2022b). In contrast, limited connectivity hinders the movement of spiders with poor dispersal ability, as they may not be able to traverse the vast distances between patches effectively (Gallé et al. 2022c). Consequently, these spiders are more likely to remain confined within their habitat patch and have limited opportunities to disperse or colonise new areas.

The differences in community composition between invaded and non-invaded patches of wild bees and predators could be because invasive species can alter the availability and quality of resources (Bartomeus et al. 2008) and habitats for these groups (Pyšek and Richardson 2007), leading to changes in their community composition. Indicator species were the most sensitive to the differences in the conditions between invaded and non-invaded habitats. We found a single indicator wild bee species for invaded patches, B. terrestris, an abundant ground-nesting bumblebee species in Europe. Our results showed that milkweed flowers attract B. terrestris. This highly mobile insect can fly over 1 km for flower resources (Osborne et al. 2008). Due to its high foraging range and sufficient tongue length to access the nectar of milkweed flowers, it is one of the few polylectic bee species that can utilise the milkweed nectar (Kephart and Theiss 2004). In invaded patches, the large quantity of nectar provided by milkweed might distract pollinating insects that are capable of utilising milkweed nectar away from native plants and affecting their pollination services of native plants (Goulson et al. 2010), with a potential negative long-term impact on them (Kovács-Hostyánszki et al. 2022).

In invaded habitats, alterations in the availability of potential spider prey associated with milkweed invasions may have cascading effects on spider populations. In our study, we found significant differences in predator community composition between invaded and non-invaded patches, with a single indicator species (P. nigrociliatus) preferring non-invaded vegetation and avoiding invaded patches. Three indicator species of invaded patches (H. bohemicus, Z. exigus, Z. segrex) were xerophilous species. The presence of milkweed provides drier and more favourable conditions for these species (Gallé et al. 2023).

Our study highlights the complex interactions between fragmentation and invasion on ecological traits of herbivores, wild bees, and predators. Our research revealed that fragmentation and invasions interacted, typically altering their respective impacts by selectively favouring opposite trait levels. Our results suggest that invasive species can significantly impact the traits of arthropods and that larger and better-connected fragments may not necessarily provide better habitats. Therefore, we suggest that restoration projects should aim to restore habitats and their connectivity and eliminate invasive species from natural habitats. This may involve targeted removal or control of invasive species while promoting the recovery of native species and their ecological interactions. Future research should focus on the effects of milkweed removal on plants and arthropods, including detailed food web analyses, assessing the direct and indirect effects of management (e.g. via altered habitat parameters) and providing an opportunity to identify species that are sensitive to or supported by management interventions. Given the complexity of interactions between fragmentation (size and connectivity) and invasion, adaptive management approaches incorporating ongoing monitoring and evaluation are crucial.