Lei Zhong

Aim: To compare the rapid shifts in body size of mainland and island populations of a native rodent and examine the mechanisms underlying these changes.Location: Thousand Island Lake, China, which was created in 1959 when the Xin'anjiang Dam was constructed for generating hydroelectricity.Taxon: The Chinese white‐bellied rat,Niviventer confucianus.Methods: Field surveys were conducted from 2015 to 2018 to collect data on body size of the rodents from a set of islands and nearby mainland sites. We constructed multiple linear models to examine the relationships between body size (length and mass) of rodents and biological variables (predators, competitors and food availability). We also conducted structural equation modelling (SEM) by constructing models via confirmatory path analysis.Results: All island populations ofN.confucianushad significantly larger body size (both body mass and body length) than their mainland counterparts. Moreover, populations on small and more isolated islands had larger body size than their relatives on big islands. The relative absence of predators (large‐bodied mammals, snakes and raptors) on islands was most strongly associated with shifts in the body size of rodents. The documented changes occurred after only a half‐century of fragmentation.Main conclusions: The observed rapid body enlargement of rodents after habitat fragmentation is consistent with a release from predation pressure. SEM indicated that island area, rather than island isolation, had positive effects on the abundance of predators, interspecific competitors and food resources, which then had an indirect impact on body size of the rodents. In this study, we report a remarkably rapid case of mammal morphological shifts in a small mammal in response to habitat fragmentation. Given the omnipresence of dams and other anthropogenic disturbances, our findings suggest that a wave of rapid phenotypic shifts in terrestrial vertebrates is taking place in the Anthropocene.

Aim: Understanding long‐term community assembly trajectories following habitat fragmentation remains a major challenge. Varying time since disturbance and heterogeneous initial starting conditions confound real‐world comparisons of re‐assembly trajectories in fragmented versus intact systems. Here, we overcome these issues in a novel manner using a system of islands and adjacent mainland habitats with common initial starting conditions following clear‐felling of forests during hydroelectric dam construction in 1959. Fifty years after resetting ‘future‐islands’ and ‘future‐mainland sites’ to common initial starting conditions, we determine the outcome of re‐assembly trajectories in plant taxonomic, phylogenetic and functional trait composition, and how these have diverged between islands versus continuous mainland habitats.Location: Thousand Island Lake, China.Methods: We surveyed woody plant communities on 26 islands and 26 adjacent mainland plots, and characterized 12 functional traits for 91 woody plant species (>46,000 total stems with DBH ≥1 cm). We used variance partitioning analysis to discriminate major environmental and spatial drivers of compositional differences, and generalised linear models to test trait and phylogenetic responses to underlying environmental gradients in island versus mainland plots.Results: We found striking differences in species composition between island and mainland plots. Variance partitioning showed that species composition on islands was governed by local environmental variables, suggesting an important role for environmental filtering, while compositional variation across mainland plots was determined by spatial variables. Ten of 12 functional traits showed significant divergence in community‐weighted mean values between island and mainland plots. Moreover, most traits exhibited significant interaction effects between isolation (island versus mainland) and underlying environmental gradients in soil fertility and stress tolerance.Main conclusion: These findings strongly suggest divergence in trait–environment relationships following isolation in fragmented habitats, with important implications for understanding future re‐assembly trajectories in fragmented systems that now dominate the planet.

Aims: With the increase of land-use change, habitat fragmentation has become a major factor affecting plant diversity. Generally, when the adult density is high, the survival rate of conspecific saplings may decline and provide more resources and space for other species to maintain a high local species diversity. Therefore, the spatial relationship of conspecific saplings and adults (SRCSA) can regulate plant diversity. However, very few studies have assessed SRCSA within fragmented forests, and we still largely unknown the relationship between SRCSA and species diversity in fragmented landscapes. Methods: We calculated the mean strength of SRCSA using the spatial distribution data of saplings and adults with a log-transformed hierarchical offset-power model on 27 islands in a land-bridge island system. The higher strength of SRCSA reflects a more negative relationship between the density of conspecific saplings and adults on an island (i.e. the looser pattern of SRCSA). We tested the relationships among island attributes (i.e. island area, the distance to the mainland—MD, and the distance to the nearest island—ND), SRCSA and species diversity. Important Findings: The strength of SRCSA increased with ND. Meanwhile, the species diversity significantly increased with the strength of SRCSA, and island area and the strength of SRCSA independently explained 26% and 6% of variation of species diversity, respectively, and their interactions explained 8%. Shade-intolerant and low-abundant species showed looser patterns of SRCSA. Our study suggests that SRCSA may have the effect of excluding species with a relatively low abundance in isolated island forests, which illustrates the importance of biotic interactions in maintaining plant diversity in fragmented forests. Our results also emphasize that we should consider forest connectivity when testing the conspecific negative density dependence.

Numerous mechanisms have been proposed to explain the island species–area relationship (ISAR), including the passive sampling effect, the habitat diversity effect, ecological drift and a group of mechanisms jointly termed environmental filtering (or abiotic selection process). However, determining which of these hypotheses underpins the ISAR in a given system remains an open and controversial question, particularly with regards to the environmental filtering. The effect of environmental filtering on the ISAR for woody plants was tested on 29 islands in a subtropical land‐bridge archipelago. Island area had a significant effect on species richness, while isolation was not found to have an effect. After controlling for the passive sampling effect and habitat diversity using a systematic sampling design, we found that a positive ISAR remained. After controlling for island area and isolation, species richness was significantly affected by environmental conditions (e.g. total phosphorus, soil depth and bulk density). When a null model was used that randomly assigned individuals to islands from the regional pool, the difference between the observed and expected richness was greater for smaller islands relative to larger islands. Synthesis. Our results point to an overlooked role of environmental filtering in driving the island species–area relationship (ISAR). Consequently, our study highlights the importance of considering deterministic process in future analyses of the ISAR.