Search

1 Forest species in West Africa show clear distribution patterns in relation to the principal rainfall gradient and associated soil fertility gradient in the region. The distribution of each species presumably reflects environmental limitation of seedling establishment, but the segregation of seasonal drought or soil conditions as controlling factors is obscured because soil nutrient availability and rainfall strongly covary. 2 Frequency of occurrence of individual species in four categories based on rainfall and a composite fertility variable are compared with the distribution of all samples. Four possible groups of species are identified, each showing a bias to one of the four site categories: dry-fertile, wet-fertile, dry-infertile and wet-infertile. A fifth group of species include those whose distribution is not significantly different from that of all sample sites. 3 Apart from the indifferent species, the wet-infertile group includes the majority of the tree species tested, including two ectomycorrhizal legumes; the wet and dry fertile groups are notable for the inclusion of deciduous Sterculiaceae. No species were found to be significantly biased to the inhospitable dry-infertile forests, which are composed principally of species with widespread distribution. 4 Experiments are proposed to test the drought tolerance and soil nutrient sensitivity of species implied by these results. Keywords: drought, plant nutrients, rain forest, West Africa

1 Soil and plant-water relations were studied in a canopy opening and surrounding forest shade in a moist tropical forest in Ghana using seedlings of two tree species with different regeneration strategies, the pioneer Terminalia superba and the non-pioneer light demander Entandrophragma utile. 2 During two consecutive dry seasons, soil matric potential varied from -30 kPa at the end of the rainy season, to below -2.5 MPa in the middle of the dry season. During part of the first dry season soil matric potential was higher in the gap than in the surrounding forest. In the second dry season no differences were observed. 3 Leaf water potential varied from about -750 kPa at dawn for unstressed seedlings of both species to < -2.5 MPa when seedlings were permanently wilted. For both species, the maximum stomatal conductance of unstressed seedlings varied from 300 to 450 mmol [m.sup.-2] [s.sup.-1] for those in the gap centre to 150 mmol [m.sup.-2] [s.sup.-1] for those under forest shade. Maximum conductance decreased to between 10 and 30 mmol [m.sup.-2] [s.sup.-1] in drought-stressed seedlings during the middle of the dry season. The two species responded similarly to a drying soil profile but the decrease in leaf water potential and stomatal conductance was more rapid under forest canopy than in the centre of the gap. 4 Only 7% of the seedlings of Terminalia superba and 4% of Entandrophragma utile survived the dry season under forest shade. In contrast, 85% and 78%, respectively, survived of those growing in the centre of the forest gap. At all positions, seedlings of T. superba had larger relative height growth rates during the wet season than E. utile (29.6, 21.4 and 8.4% [month.sup.-1] compared with 5.2, 20.0 and 1.6% [month.sup.-1], for gap centre, margin and forest shade, respectively). 5 The observed changes in leaf water potential and stomatal conductance indicated that despite the reduced irradiance, shaded seedlings experienced more drought stress during the dry season than seedlings growing in the forest gap. The reasons for this and possible consequences for trees which differ in leaf phenology and regeneration strategy are discussed. Keywords: forest gaps, leaf water potential, seasonal droughts, soil matric potential, stomatal conductance

1 Milicia regia (Moraceae) is a dioecious light-demanding (Pioneer) timber tree of West Africa. Experiments in shadehouses were used to examine the influence of light and soil source (beneath and away from conspecific adults) on mortality and growth of its seedlings in relation to fungal pathogens and arbuscular mycorrhizal (AM) colonization. 2 The experiment in Ghana included treatments for light (2% and 20% of unshaded irradiance), three soil sources (under female, under male and 200 m distant from adults), and two different fungicide treatments. 3 The results showed significant effects of irradiance, soil source and fungicide on both seedling mortality and accumulated biomass. Biomass was greater and mortality reduced at higher irradiance. Fungicide application reduced mortality in 2% irradiance, but had no significant effects at 20%, although AM infection of seedlings grown in soils from beneath adults was severely reduced. Mortality was greatest in soil from beneath female trees and least in soil distant from adult conspecifics. However, biomass in 20% irradiance was greater in soil from beneath females, possibly because of enhanced AM colonization: seedlings had a significantly higher percentage root length colonized by AM fungi beneath adult conspecifics, and biomass accumulation was significantly correlated with root length colonized by AM fungi. 4 There were significant interactions amongst light, soil source and fungicide treatments that attest to the complexity of spatial effects on seedling establishment in tropical forest. Whilst there was clear evidence of pathogen-mediated effects that provides support for Janzen-Connell spacing mechanisms, and escape from the parent tree increases the chances of arriving in gaps, spatial differences in AM fungal communities may act in the opposite direction and enhance seedling growth close to adult conspecifics. 5 The complexity of interactions between abiotic and biotic factors and the way in which these factors can vary spatially to affect seedling recruitment may be an important factor contributing to the maintenance of tree species diversity in tropical forests.