17Land degradation results in declining biodiversity and disruption of ecosystem 18 functioning worldwide, particularly in the tropics 1 (Nectariniidae, Pycnonotidae), and three lizard species (Gekkonidae, Scincidae). 95Restoration markedly changed pollinator numbers, behaviour, performance, and 96 network structure in inselberg communities. Six to 14 months after restoration, 97 number of pollinator species was on average 21.6% higher across the four restored 98 compared to the unrestored inselbergs (Fig. 2). Monthly pollination networks showed (Fig. 2, Table1). Overall, restored networks were more generalised than 103 unrestored networks (H 2 ʹ, Fig. 2). 105The observed network responses to restoration were mirrored by the plant The observed changes in pollinator interaction behaviour and network structure had Fig. 2), which correlated with a larger total 138 fruit production (fruit crop) and higher fruit set (proportion of flowers producing fruit) 139 across the most common species (Fig. 3, Table 1). The three endemic palms (Fig. 3). Plants at restored sites had similar fruit set levels throughout the range

Summary1. Invasive alien plant species pose a severe threat to native plant communities world-wide, especially on islands. While many studies focus on the direct impact of alien plants on native systems, indirect effects of plant invaders on co-flowering natives, for example through competition for pollination services, are less well studied and the results are variable. 2. We used six temporally and taxonomically highly resolved plant-pollinator networks from the island of Mahe´, Seychelles, to investigate the indirect impact of invasive alien plant species on remnant native plant communities mediated by shared pollinators. We employed fully quantitative network parameters and information on plant reproductive success, and pollinator diversity and behaviour, to detect changes in plant-pollinator networks along an invasion gradient. 3. The number of visits to and fruit set of native plants did not change with invasion intensity. Weighted plant linkage and interaction evenness, however, was lower at invaded sites than at less invaded sites. These patterns were primarily driven by shifts in interactions of the most common pollinator, the introduced honey bee Apis mellifera, while weak interactions and strong native interactions remained unchanged. 4. Synthesis. The implications of these findings are twofold: first, quantitative network parameters are important tools for detecting underlying biological patterns. Secondly, alien plants and pollinators may play a greater role in shaping network structure at high than low levels of invasion. We could not show, however, whether the presence of invasive plants result in a simplification of plant-pollinator networks that is detrimental to native plants and pollinators alike.

Habitat fragmentation and changed land use have seriously reduced population size in many tropical forest tree species. Formerly widespread species with limited gene flow may be particularly vulnerable to the negative genetic effects of forest fragmentation and small population size. Vateriopsis seychellarum (Dipterocarpaceae) is a formerly widespread canopy tree of the Seychelles, but is now reduced to 132 adult individuals distributed in eleven sites. Using ten microsatellite loci, a genetic inventory of all adult trees and a sample of 317 progeny, we demonstrate that despite its restricted range, overall genetic diversity was relatively high (H(E) : 0.56). The juvenile cohort, however, had significantly lower allelic richness (adults R(S) : 3.91; juveniles R(S) : 2.83) and observed heterozygosity than adult trees (adults H(O) : 0.62; juveniles H(O) : 0.48). Rare alleles were fewer and kinship between individuals was stronger in juveniles. Significant fine-scale spatial genetic structure was observed in remnant adults, and parentage analysis indicated that more than 90% of sampled progeny disperse <25 m and pollen dispersed <50 m. The molecular data confirmed that two populations were derived entirely from self-fertilized offspring from a single surviving mother tree. These populations produce viable offspring. Despite this extreme genetic bottleneck, self-compatibility may provide V. seychellarum with some resistance to the genetic consequences of habitat fragmentation, at least in the short term. We discuss our findings in the context of other rare and threatened dipterocarp species which are vulnerable to miss-management of genetic resources and population fragmentation.

No abstract

Questions Which physical or chemical removal method is most effective in controlling woody invasive alien plant (IAP) species in sensitive conservation areas, while minimizing the impact on different life stages of native plant communities? Is there a distinct influence of removal method on successional trajectories of post‐treatment seedling communities? Location Mid‐altitude inselbergs, Mahé, Seychelles, Indian Ocean. Methods We conducted two field experiments. First, to test the effects of four IAP control methods (cut, 5%, 10%, 15% Tordon 101 herbicide concentration) on ten woody IAP species, we treated a total of 320 adults and monitored mortality for 18 wk. Second, we set up five blocks each with four plots (4 m × 4 m) at three inselberg sites (N = 60 plots) and recorded the response of seedlings, saplings and adults of all native and IAP species to four Tordon 101 (a.i. picloram and 2,4‐D) herbicide and manual control treatments (cut, cut‐stump, foliar spray, control) over 22 mo. Treatment effects were analysed with GLMM models, and community responses were investigated with NMDS and multivariate ANOVA. Results The first experiment showed that IAP mortality (98.3%) was similar across all three herbicide concentrations, but markedly higher than mortality of cut plants (42.5%). In the second experiment, abundance, but not diversity, of native seedlings declined with herbicide application in the first 6 mo after treatment. Ordination analysis showed that seedling communities experienced high turnover driven by early successional IAP species and abundance, and that between‐plot variation was determined by site‐specific differences, with only a weak treatment effect. Foliar spray resulted in high mortality of native saplings (52.7%) and adults (18.3%), and both herbicide applications reduced adult growth over a post‐treatment period of 12 mo. Conclusion IAP removal controlled plant invasion processes and altered community composition in the short term, regardless of removal method. Post‐treatment trajectories of seedling communities are determined by site‐specific differences. Chemicals can strongly increase the efficacy of IAP control, but short‐term detrimental effects on native saplings and adults may be expected. As re‐invasion by IAPs is likely, a long‐term restoration plan is required. Management strategies that aim to conserve β‐diversity should employ a landscape approach to habitat restoration to buffer against site idiosyncrasies.