The red land crab (Gecarcoidea nalalis) is the major collector of fruits and seeds from the rain forest Moor on Christmas Island, Indian Ocean. At three sites, they rapidly removed fruits and seeds of 17 species tested (usually within 12 h after placement) and most were taken to crab burrows. Direct observations of seed predation, and application of estimated chelar forces of red crabs indicated that seeds of many species, especially those with thin seed coats, are likely to be killed through handling. However, removal of some resistant seeds to burrows may be favourable for seedling establishment when canopy gaps occur. Fruit removal rates by crabs were positively related to nitrogen concentration in fruit tissues but negatively correlated with condensed tannins. These observations suggest that seed predation by land crabs may affect the success of colonization and the relative abundance of plants on oceanic islands.

Soil mineralization and nitrification rates were measured in (1) undisturbed and felled gaps of varying size in terra firme forest, and (2) along a forest – savanna transect, on Ilha de Maracá in northern Brazil. Both rates were similar to those found in studies of other forests with a marked seasonal rainfall pattern. However, rates were much lower than those of tropical forests where there is little seasonality in rainfall. A major finding was that the highest rates were during the transition between dry and wet seasons, implying that wetting and drying may be an important initiator of soil nitrogen flux. Felling had little effect on either process up to a gap size of 2500 m2. In the forest-savanna study, nitrogen mineralization was lower in the savanna in all seasons, but in the wet season when the savanna soils were water-logged NH4+-N was immobilized. Experimental additions of nutrients identified two important results; first that added NH4+-N was quickly immobilized rather than nitrified – this may be an important nutrient conservation mechanism, and second that soil calcium appeared to be a limiting factor, either directly or through a pH effect.

The germination response of Uraria picta to five environmental factors was investigated. It germinated both in light and dark but germination in the light was significantly better than in the dark; there was further germination when seeds first placed in the dark were later brought into the light. It germinated well in the three soil types used but germination in humus and sand was significantly better than in red earth. There was an increase in germination with an increase in burial up to 1.5 cm, thereafter germination levelled off. There was no germination in water-logged conditions and there was an increase in germination as soil moisture decreased. Alternating temperatures of 31/21°C and 31/15°C favoured germination while 21/15°C inhibited germination. The results are discussed in relation to the phenology and distribution of the species.