The island rule- a general pattern of dwarfism in large species to gigantism in small species on islands relative to mainland- is typically seen as a macroevolutionary phenomenon. However, it remains unknown whether the ecological processes associated with abiotic and biotic factors generate a pattern of plant size variation similar to the island rule. Through measuring plant height for 29623 individuals of 50 common woody species in the Zhoushan Archipelago (8500 years old and yet to undergo major evolutionary adaptation) and the adjacent mainlands in Eastern China, we examined whether island area and remoteness, resource availability, environmental stress, plant-plant competition and insect herbivory can explain the pattern of plant size variation. We found pronounced variations in plant height, similar to those of the island rule. Further analyses revealed that islands with low resource availability, such as low soil organic matter content and low precipitation, had a high degree of dwarfism; islands experiencing high environmental stress, such as high soil pH, had a high degree of dwarfism; and islands experiencing less plant-plant competition had a high degree of gigantism. The magnitude of plant dwarfism was also higher on small and remote islands than on larger and nearer islands. These results highlight the importance of ecological processes associated with abiotic and biotic conditions in shaping the island rule-like patterns of plant size variation. Our study therefore suggests that the island rule can be caused by both ecological and evolutionary processes. Given that the age of our studied archipelago is too young to undergo major evolution, our results evidenced that ecological processes likely played a prominent role for generating the island rule-like patterns. Future studies on the island rule need to perform experiments to disentangle evolutionary from ecological mechanisms.

Forest litter decomposition is considered as an essential ecosystem process affecting carbon and nutrient cycling in mountains. However, there exists high uncertainty in accurately estimating the contribution of litter decomposition to terrestrial ecosystems, largely due to the incomparability of different studies and the data limitation in microclimate and non-climatic factors at spatially matched scales. Here we used the tea bag index (TBI) as a standardized protocol to evaluate spatial variations in forest litter decomposition rate (k) and stabilization factor (S) across 10 mountains spanning a wide range of subtropical and tropical forests. Based on the coordinated experiment of 6,864 teabags in 568 sampling sites along elevations, we evaluated the importance of 10 environmental factors covering soil microclimate, edaphic properties, plant diversity, and topography on k and S by using model averaging and linear-mixed effects models. Of the 10 mountains, we found a consistently decreasing pattern of k and an increasing pattern for S along elevations. And the significant effect of k with elevation was mainly found in the western and northmost mountains, while the effect of S occurred in the western and southernmost mountains. For microclimate, there was a general importance of soil temperature (coef. = 0.48) and temperature variation in the growing season (coef. = 0.36) in k, and soil temperature (coef. = -0.46) and moisture variation on S (coef. = -0.36). The dominant role of soil microclimate was mainly found in western mountains with relatively cold environments. For non-climatic drivers, a significant effect of tree diversity on k and a negative correlation of edaphic and topography with S in the western and southern mountains were detected. These findings provide a general understanding of spatial variations of driving factors in forest litter decomposition and highlight a dominant control of soil microclimate in cold forests in high elevations and latitudes.

Soil phosphorus (P) is an important nutrient limiting plant productivity. The elevational pattern of soil P concentrations is widely used to indicate how P pools are affected by the climate. Despite previous research finding various patterns of soil P concentration across different elevations, little research has been conducted on how bedrock variation in an elevational transect affects P pools. In this study, we examined the elevational patterns of soil P pools (e.g., labile inorganic P (Pi) and organic P (Po), moderately labile Pi and Po, primary mineral P, and occluded P) in top-soils in an elevational transect with alternating bedrock types of clasolite and limestone in a subtropical karst forest. The results showed that concentrations of soil total P, labile Pi, primary P, and occluded P were significantly higher in limestone soils than clasolite soils, while there were no significant differences in labile Po, moderately labile Pi and Po concentrations between the two bedrocks. There were generally no significant linear elevational trends for soil P pools. Our results revealed that the bedrock type had a greater influence on soil P pools than climatic factors. Our study emphasizes the importance of using bedrock changes as a driving factor for the spatial distribution of soil phosphorus pools in mountainous ecosystems.