The research, published in the Journal of Hymenoptera Research, set out to test whether the spread of Canadian goldenrod affects the abundance, diversity, and species composition of ants in meadows with contrasting exposures. Ants are among the most successful insect groups on Earth, often accounting for the largest share of biomass in terrestrial invertebrate communities. By constantly building nests and tunnels, they act as soil engineers, a process known as bioturbation that turns over massive amounts of soil and enhances aeration, water infiltration, and nutrient availability for plants. Many plant species have evolved seeds with attractive elaiosomes specifically for ants, which consume the elaiosome and leave the intact seed in a sheltered, nutrient-rich location within the nest, making ants key dispersers of seeds in forests and scrublands. Ants are also highly efficient cooperative predators, controlling other invertebrate populations by preying on the eggs and larvae of various pest species. Any invasive plant that disrupts ant communities could therefore ripple through entire ecosystems.

Previous studies from Poland and elsewhere have painted a grim picture for ants facing goldenrod invasion. Research has shown that the spread of Solidago canadensis decreased nest density and ant species abundance compared to non-invaded meadows, and also changed the dominant ant species in the area. The genus Myrmica was strongly affected by the presence of Canadian goldenrod, while other species such as Lasius niger were stimulated in the invaded sites. In the case of Myrmica rubra, it was shown that most of their energy is used for gyne production and less for workers, as an adaptation to an environment in which goldenrods spread, to extend to a location with more advantageous conditions. The spread of Canadian goldenrod was associated with lower ant diversity and even reduced numbers of anthills built in smaller sizes than normal. Ant communities were also affected indirectly by a lack of aphids on the surface of Solidago canadensis, giving ants an additional reason to avoid this species.

To investigate whether these patterns hold in central Romania, the research team carried out field work during August 2018 on the Târnava Plateau in Transylvania, near the regions of Daneș, Pelișor, Bârghiș, and Biertan. The study employed a paired design with six sites, each containing one invaded and one non-invaded meadow, for a total of twelve sampling locations. The geological substratum is composed mainly of Pannonian sedimentary deposits including clays, sands, and poorly cemented grits, as well as Sarmatian, isolated Badenian, and Quaternary deposits found along the valleys. The floristic inventory of the whole plateau consists of 897 vascular taxa distributed in 100 families, within which vulnerable plant species form vegetation associations such as Festuco-Brometea, Rhamno-Prunetea, and Querco-Fagetea. The meadows were chosen based on comparable vegetation and being spatially paired as invaded and non-invaded in each site.

The experimental design included two types of exposure: northern and southern. The sampling meadows in the Biertan region had northern exposure, whereas those in the other investigated regions had southern exposure. Ant sampling was carried out using Barber pitfall traps, a well-suited technique for collecting ground-dwelling invertebrates. Glass or plastic containers of 10 centimeters in diameter were buried in the ground so that the opening surface was flush with the soil surface. A roof 4 to 6 centimeters higher was installed above the traps to avoid filling them with water from precipitation. The traps were half-filled with a preservative liquid of glycol and water in a 1:1 solution and collected after 14 days. A total of 180 Barber traps were installed across the 12 sites, with each sampling site consisting of 15 pitfall traps arranged in a 3 by 5 network with 5 meters between traps. For comparability, sampling was carried out in patches where at least 49 square meters were covered with at least 75 percent Solidago canadensis for the invaded plots, and a similar surface was used for non-invaded plots where the plant was absent.

The myrmecological material was analysed and identified to species level using established taxonomic keys. Statistical analysis was based on 162 of the originally placed 180 traps, as some were destroyed. The researchers employed Generalized Linear Mixed Models to test the main and interactive effects of exposure and invasion status on ant abundance and species richness per trap, controlling for the non-independence of samples collected from each site by including the sampled sites as a random intercept. Diversity analysis in terms of species richness and the Shannon–Wiener index was carried out at the meadow level, standardized using sample-size-based rarefaction and extrapolation curves to 15 traps in each meadow. Principal Component Analysis was performed on the Hellinger-transformed species abundance matrix to visualize gradients in ant community structure, and Redundancy Analysis was used to test whether variance in species composition could be explained by exposure and invasion status.

Altogether, the team identified 21 ant species belonging to four subfamilies, all common in Romania. In terms of zoogeographical elements, Euro-West-Siberian and South-Palearctic species were dominant, whose specific habitats are dry meadows and dry habitats. Comparing ant species number per trap between invaded and non-invaded habitats, the researchers found no significant differences, although a slightly higher number of species was observed in the non-invaded sites. Neither the main effect of invasion status nor exposure significantly influenced species richness per trap, and the interaction effect was also not significant. For ant abundance, there was no statistically significant main effect of invasion status, indicating that Solidago canadensis invasion did not reduce ant abundance across all sites at the fine scale, although there was a trend of slightly increased ant abundances in the non-invaded sites.

Interestingly, there was a marginally significant effect of exposure on ant abundance, with a trend towards lower abundance on southern slopes compared to northern slopes. The northern exposure was characterized by higher abundances compared to the southern exposure. At the meadow level, the estimated species richness of ant communities did not differ significantly based on the presence of the invasive species or the exposure of the site, probably due to the small number of replicates and large variance, though the median species richness per meadow was about twice as high on northern slopes as on southern slopes. The Linear Mixed Model for the Exponential Shannon Index showed no significant main effects of invasion status or exposure, and the variance due to exposure was much lower, suggesting that only rare species contributed to the increased species richness on the northern slope.

The community composition analysis revealed that exposure, not invasion, was the primary driver of ant assemblage structure. The first two axes of the Principal Component Analysis explained approximately 48.6 percent of the variability in ant species composition. Species such as Lasius niger, Formica cunicularia, Lasius alienus, and Formica rufibarbis occurred mainly in both invaded and non-invaded habitats but with a preference for northern exposure, whereas species such as Myrmica scabrinodis, Solenopsis fugax, Myrmica rubra, and Lasius flavus occurred mainly in southern-exposed meadows. In the constrained Redundancy Analysis, exposure explained 26.8 percent of the observed variation in ant community composition and correlated significantly with the first axis. Species including Lasius alienus, Polyergus rufescens, Formica rufibarbis, Camponotus aethiops, Camponotus piceus, and Myrmica gallienii were predominantly found in habitats with northern exposure, such as those in the Biertan sites, while Tapinoma erraticum, Lasius flavus, Solenopsis fugax, Myrmica scabrinodis, Ponera coarctata, and Myrmica sabuleti were more frequently found in habitats with southern exposure. In contrast, invasion had no significant effect on species composition, although there was an indication that Tetramorium cf. caespitum increased in most of the invaded sites.

The researchers propose several explanations for why their findings contrast with earlier studies showing negative effects of goldenrod on ants. All the sites with northern exposure in the study, located in the Biertan region, are managed by traditional mowing at least two times per year. Thus, the structure of vegetation can shift and may enhance abundance of ants in these habitats compared to less frequently mowed areas. Other studies have shown that the highest positive effect on the re-establishment of native plant species was recorded with two mowing events per year in meadows invaded by Solidago. Slope aspect is a well-established driver of microclimatic variation in temperate ecosystems, with southern exposures typically experiencing higher solar insolation and soil temperatures compared to northern slopes. However, this expected thermal advantage did not translate into significant differences in ant abundance, possibly due to vegetative buffering or high intra-meadow heterogeneity. In harsh environments such as invaded habitats, cooler and more humid north-facing slopes may act as microrefugia, protecting ants from lethal temperatures and desiccation, and often supporting unique, functionally specific ant assemblages that could not survive the harsh xeric conditions of intensive heat typical of southern slopes. The authors acknowledge that since the northern-exposure meadows were all situated in the Biertan region with more frequent mowing, they cannot fully distinguish the effects of management from regional microhabitat effects. Nevertheless, the study concludes that the invasive plant Solidago canadensis causes rather a limited effect on the diversity and abundance of ant communities in central Romania, while environmental factors and the overall landscape must be included in community ecology analyses of these systems.

Subject of Research: Ecological effects of Solidago canadensis invasion on ant communities in central Romanian meadows

Article Title: The invasion of Goldenrod ( Solidago canadensis ) does not affect the ant communities ( Hymenoptera , Formicidae ) in central Romania meadows

Article References: Tăușan, I., Cocîrlea, M. D., & Oancea, S. (2026). The invasion of Goldenrod (Solidago canadensis) does not affect the ant communities (Hymenoptera, Formicidae) in central Romania meadows. Journal of Hymenoptera Research, 99, 231-248. https://doi.org/10.3897/jhr.99.158519

Image Credits: AI Generated

DOI: 10.3897/jhr.99.158519

Keywords: invasion, Goldenrod, Solidago, canadensis, does, affect, communities, Hymenoptera, Formicidae, central, Romania, meadows