This study expands the limited knowledge of the frequency of cyanogenesis in natural plant communities, includes novel reports of cyanogenesis among a range of taxa and characterizes patterns in intra-plant and intra-population variation of cyanogensis.

Our ability to model global carbon fluxes depends on understanding how terrestrial carbon stocks respond to varying environmental conditions. Tropical forests contain the bulk of the biosphere's carbon. However, there is a lack of consensus as to how gradients in environmental conditions affect tropical forest carbon. Papua New Guinea (PNG) lies within one of the largest areas of contiguous tropical forest and is characterized by environmental gradients driven by altitude; yet, the region has been grossly understudied. Here, we present the first field assessment of aboveground biomass (AGB) across three main forest types of PNG using 193 plots stratified across 3,100-m elevation gradient. Unexpectedly, AGB had no direct relationship to rainfall, temperature, soil, or topography. Instead, natural disturbances explained most variation in AGB. While large trees (diameter at breast height > 50 cm) drove altitudinal patterns of AGB, resulting in a major peak in AGB (2,200-3,100 m) and some of the most carbon-rich forests at these altitudes anywhere. Large trees were correlated to a set of climatic variables following a hump-shaped curve. The set of "optimal" climatic conditions found in montane cloud forests is similar to that of maritime temperate areas that harbor the largest trees in the world: high ratio of precipitation to evapotranspiration (2.8), moderate mean annual temperature (13.7°C), and low intra-annual temperature range (7.5°C). At extreme altitudes (2,800-3,100 m), where tree diversity elsewhere is usually low and large trees are generally rare or absent, specimens from 18 families had girths>70 cm diameter and maximum heights 20-41 m. These findings indicate that simple AGB-climate-edaphic models may not be suitable for estimating carbon storage in forests where optimal climate niches exist. Our study, conducted in a very remote area, suggests that tropical montane forests may contain greater AGB than previously thought and the importance of securing their future under a changing climate is therefore enhanced.

The role of fire in governing rainforest-eucalypt forest ecotone dynamics is of theoretical interest and has conservation management implications. Several eucalypt forests in the Wet Tropics of Australia have an endangered status due to extensive conversion to rainforest. Rainforest plants are known to survive occasional low intensity fires in the eucalypt forest ecotone. However, the ability of rainforest plants to survive frequent fires remains untested. The timing of rainforest expansion is also a subject of interest, and is generally considered to be delayed until fire has been absent for several years. We used 14 years of data collected across 13 plots in the Wet Tropics of north-eastern Australia to test predictions regarding rainforest seedling recruitment and post-fire regenerative capacity.The 13 plots received different numbers of fires, between zero and five, over the 14-year study. The recruitment of new rainforest plants in the ecotone was most abundant in the initial year after fire. If this post-fire pulse of recruitment is left undisturbed, it can facilitate the subsequent germination of additional rainforest species. The removal of grass cover, whether temporarily in the immediate post-fire environment or once a developing rainforest mid strata shades out grasses, appears crucial to abundant rainforest recruitment. A variety of tropical rainforest species can persist under a frequent fire regime through resprouting. The difference in the mode of resprouting, between ground-level coppicing rainforest plants and canopy resprouting eucalypt forest trees, is the critical mechanism that causes regular fire to maintain an open structure in eucalypt forests. The inability of rainforest species to maintain their height when fires fully scorch their crowns, temporarily resets the forest's open structure and delays the rainforest's ability to dominate through shading out grasses to transform the ecosystem into a closed forest.

We examined the impact of severe cyclone 'Larry' on the vegetation structure of monoculture and mixed species timber plantations, restoration plantings and reference sites in upland rainforests on the Atherton Tableland, north Queensland, Australia. Sites were initially assessed in 2000 and resurveyed in 2006, 6-8 months after the cyclone traversed the region. In both surveys, timber plantations had a relatively open canopy, grassy understorey and few shrubs or small-sized trees; whereas restoration plantings had a relatively closed canopy, an understorey of bare ground, leaf litter and rainforest seedlings, a high density of small-diameter trees and a moderate representation of special life forms characteristic of rainforest. Cyclone damage varied with tree size, site type, proximity to the cyclone and stem density. First, the proportion of trees that were severely damaged by the cyclone (major branches broken, stem snapped or pushed over) increased with the diameter of trees across all site types. Second, damage to larger-sized trees (>10 cm d.b.h., >20 cm d.b.h.) was proportionally highest in monoculture plantations, intermediate in mixed species plantations and rainforest, and lowest in restoration plantings. Third, within site types, damage levels decreased with distance from the cyclone track and with stem density.There was no evidence that topographical position influenced damage levels, at least for timber plantations.We tentatively attribute the high levels of damage experienced by timber plantations to their relatively open structure and the large size of stems in plantations. Restoration plantings generally escaped severe damage by the cyclone, but their continued development towards rainforest conditions may require a coordinated monitoring and maintenance programme to address the potential threat of weed invasion.

Context There is growing evidence that vertebrates inhabiting the extensive savannas of northern Australia are undergoing a widespread decline as a result of the effects of anthropogenic land management such as the grazing of domestic stock. Despite the ubiquity of pastoral grazing in the Australian savannas, few studies have examined the changes in terrestrial vertebrate fauna following destocking. Aims The present study monitored the response of birds, mammals and reptiles to destocking of a newly acquired conservation reserve in north-eastern Australia. Methods The vertebrate fauna was sampled over a 5-year period. Standardised 1-ha survey was conducted twice a year in 2006, 2007 and 2010, at 40 sites representing six habitat types. Key results The fauna assemblage, the abundance and richness of major taxa, and the abundance of a suite of individual species were found to vary significantly with time since destocking. Although some of the observed trends were consistent with previously reported responses of vertebrates to grazing, in general species richness and abundance did not increase linearly over time since destocking, with an overall decline in the first year, and an increase in the subsequent survey. Mammals remained at very low abundance and displayed a trend contrary to that for birds and reptiles, and variation was often confounded by habitat type. Conclusions In general, where there has been a long history of pastoral land management, destocking alone may not induce short-term increases in the vertebrate fauna thought to be affected by grazing in Australian savannas. Implications Monitoring the outcomes of conservation management activity is a critical component of understanding the success, failures and adaptation needed to maximise the costs and benefits of conservation investment. The recovery of the vertebrate fauna thought to be of conservation concern in relatively intensively used, long-grazed landscapes may be lengthy and contingent on other factors, such as periods of favourable weather, or understanding the interactive effects of herbivore removal, fire pattern and feral predators. In such landscapes, it is possible that recovery of some elements of the vertebrate fauna may not occur without deliberate interventions, such as reintroductions or intense predator control.