Seedlings of Pentaclethra macroloba (Willd.) Kuntze, a dominant, shade-tolerant species of tropical moist forest in Costa Rica, were raised under three light conditions to assess their acclimatization and growth responses to irradiance. The light environments used were full sun, partial shade (25% of full sun), and full shade (1% of full sun). To test the effect of a sudden change in light level such as would occur in the event of a treefall gap, the seedlings in the above conditions were switched between environments after two months and grown for an additional 1.5 months. Plants in full sun and partial shade had similar total weight; switching between the two environments had no effect on biomass. Plants switched from full sun and partial shade to full shade had negative growth as a result of negative CO2, flux and leaf abscission. Plants transferred from full shade to full sun had lower growth rates than those switched from full shade to partial shade because of severe leaf damage in full sun. The previous environment significantly affected the response of most growth and size characteristics to the present environment. Maximum photosynthesis changed only 30% in response to light level during growth. However, large changes in other photosynthetic and structural characteristics were found. Leaves developed in full shade had lower respiration, leaf thickness, and stomatal density and higher apparent quantum yield, specific leaf area, and chlorophyll content than leaves from full sun. Osmotic potentials were similar between treatments. Despite only small changes in maximum photosynthesis, acclimatization to extreme shade or sun rendered leaves unsuitable for large changes in environment. Consequently, in the event of a large treefall gap, plants already present in the understory will require a substantial period of adjustment before they respond to the increase in light.

The genetic diversity of Myrciaria floribunda, a common Atlantic Forest tree, was investigated in six populations located in two small fragments (10 and 18 ha), two medium-sized fragments (36 and 44 ha) and one large fragment (3003 ha). Two populations occur in the large fragment. It is expected that smaller fragments should have lower genetic diversity and higher inbreeding. Distances between fragments varied from 0.66 to 10 km. On average 32 young trees smaller than 20 cm basal girth were sampled in each population. Allozyme electrophoresis was carried out, and six loci were scored. The effective number of alleles was lower for populations of the two small fragments (1.46 and 1.51) and higher for populations of the large (1.62 and 1.71) and medium ones (1.69 and 1.84). Small fragments showed lower values of expected and observed heterozygosities than large and medium fragments. Most of the genetic variability occurs within populations, and there was a moderate genetic variation among them ($\hat{\theta}_p$ = 0.097). Our findings show a tendency of lower genetic diversity within small and isolated fragments and higher genetic differentiation among them. But, few correlations between genetic diversity indices and fragment features (size and isolation) were significant.