Summary 1.Invasive plants can have a major impact on local plant and animal communities. However, effects of plant invasions on arthropod communities and the potential drivers have rarely been studied. 2. We present a meta-analysis of 56 studies on the impact of plant invasions on abundance and richness of local arthropod communities. Moreover, we study the role of five invader and habitat attributes to assess their influence on the direction and magnitude of effect on arthropod communities: the time since introduction; woody vs. herbaceous invaders; presence of native congeners; canopy cover of the invader; and single vs. multiple invaders. 3. We found that overall invaded habitats had a 29% lower arthropod abundance and a 17% lower taxonomic richness compared with non-invaded habitats. Woody invaders had a stronger negative impact on arthropod communities than herbaceous invaders, reducing abundance and richness by as much as 55% and 21%, respectively. 4. Synthesis. Our study demonstrates that arthropod communities are negatively affected by plant invasions, which may have substantial effects on other ecosystem features, such as pollination, food web dynamics, decomposition as well as habitat heterogeneity. Loss of arthropod diversity is generally directly associated with loss of plant species richness. Therefore, the reduction we see could be causally connected to the effect of the invader on the habitat. The physical dominance of woody invaders compared with herbaceous invaders could be a main driver for this effect.

Endemic genera on oceanic islands often evolved striking morphological and ecological differences among species, with weak postzygotic reproductive isolation. Human activities can lead to increased connectivity and can thereby promote secondary contact and hybridization between previously isolated species. We studied this phenomenon in three species of the genus Pericallis (Asteraceae) on Tenerife, Canary Islands. A total of 53 populations was sampled, including morphologically uniform as well as mixed populations containing morphologically diverse individuals. All plants were analyzed both genetically using AFLP markers and morphometrically. As expected, morphological analysis clearly separated the plants from the uniform populations in three clusters, with distinguishing characters corresponding to those used in species identification. The three species were also grouped into distinct genetic clusters in a STRUCTURE analysis, although no private alleles were observed. Adding the mixed population data to the analyses provided evidence for extensive hybridization among species, predominantly between P. cruenta and P. echinata, and between P. cruenta and P. tussilaginis, and morphological and genetic signals were congruent. The mixed populations were geographically located in between the uniform populations and were found in road verges significantly more often than the uniform populations. The observed distribution strongly suggests that hybridization between originally isolated species is recent, promoted by secondary contact due to human disturbance and the construction of roads.

No abstract

Greenhouses are a well-accepted containment strategy to grow and study genetically modified plants (GM) before release into the environment. Various containment levels are requested by national regulations to minimize GM pollen escape. We tested the amount of pollen escaping from a standard greenhouse, which can be used for EU containment classes 1 and 2. More specifically, we investigated the hypothesis whether pollen escape could be minimized by insect-proof netting in front of the roof windows, since the turbulent airflow around the mesh wiring could avoid pollen from escaping. We studied the pollen flow out of greenhouses with and without insect netting of two non-transgenic crops, Ryegrass (Loliummultiflorum) and Corn (Zea Mays). Pollen flow was assessed with Rotorod® pollen samplers positioned inside and outside the greenhouse’ roof windows. A significant proportion of airborne pollen inside the greenhouse leaves through roof windows. Moreover, the lighter pollen of Lolium escaped more readily than the heavier pollen of Maize. In contrast to our expectations, we did not identify any reduction in pollen flow with insect netting in front of open windows, even under induced airflow conditions. We conclude that insect netting, often present by default in greenhouses, is not effective in preventing pollen escape from greenhouses of wind-pollinated plants for containment classes 1 or 2. Further research would be needed to investigate whether other alternative strategies, including biotic ones, are more effective.Electronic supplementary materialThe online version of this article (doi:10.1007/s10453-011-9237-8) contains supplementary material, which is available to authorized users.