We used a field experiment to quantify the independent effects of variation in several fruit characteristics and in wind speed on dispersal of the windblown samaras of the tropical tree, Tachigalia versicolor. On Barro Colorado Island, Panama, we simultaneously released 15 different types of artificial models of these fruits, varying in seed mass, fruit area, and basic morphology, from the top of a 40—m tower, dropping them under five different wind conditions. In general, the horizontal distance a model fruit moves from the base of the tower is proportional to the average wind speed to which it was subject during the whole of its descent, divided by the square root of its wing—loading (wing—loading is the force of gravity on the fruit divided by its projected surface area). This relation was observed for fruits with masses from 1.0 to 1.9 g, with areas from 9 to 76 cm2, and with (wing—loading)1/2 from 15.3 to 30.6 mPa. Two separate changes in the design of the fruit, but with no change in the wing—loading value from that of T. versicolor, caused the models to have either a more rapid rate of descent despite the appropriate aerodynamic motion, or an improper aerodynamic motion resulting in a pronounced increase in rate of descent. Model fruits that had the lowest wing—loading (achieved by decreasing seed mass or increasing fruit area) or that were blown by the strongest mean winds were dispersed the greatest mean distances and were scattered most evenly and over the widest area. A change in wing—loading or wind speed changed dispersal distance the least and the evenness of the distribution the most. Variation in wind speed has the potential to override the effect of variation in mass or area on dispersal, unless the differences in mass or area among models are quite large.

A total of 52,467 wind-dispersed seeds from 14 tree and 32 liana species fell into 1720 seed traps in 43 paired light-gap and adjacent forest sites on Barro Colorado Island, Panama. Summed at the community level, many more wind-dispersed seeds were collected from light-gaps (61%) than from forest sites (39%). They accumulated from March through May, 1984 to a density of 328 m-2 in gaps and 207 m-2 in forest sites. In contrast, only 33% of the total of 2782 non-wind-dispersed seeds were collected in gaps. Due to the extreme heterogeneity of the seed rain, these differences between gap and forest sites were not statistically significant at the community-level. Gap sites received more wind-dispersed seeds than adjacent forest sites in only 20 of 43 locations and in 13 of 20 species, especially those with individuals of high fecundity near gap sites. Of the estimated 105 million wind-dispersed seeds contributing to the seed rain of the 50 ha study plot, only 4.1% were dispersed to the rare gap sites that enhance the establishment and growth of seedlings for many of these species.

Multilocus DNA fingerprinting methods have been used extensively to address genetic issues in wildlife populations. Hypotheses concerning population subdivision and differing levels of diversity can be addressed through the use of the similarity index (S), a band-sharing coefficient, and many researchers construct hypothesis tests with S based on the work of Lynch. It is shown in the present study, through mathematical analysis and through simulations, that estimates of the variance of a mean S based on Lynch's work are downwardly biased. An unbiased alternative is presented and mathematically justified. It is shown further, however, that even when the bias in Lynch's estimator is corrected, the estimator is highly imprecise compared with estimates based on an alternative approach such as 'parametric bootstrapping' of allele frequencies. Also discussed are permutation tests and their construction given the interdependence of Ss which share individuals. A simulation illustrates how some published misuses of these tests can lead to incorrect conclusions in hypothesis testing.

The distribution of wind‐dispersed seeds around a parent tree depends on diaspore and tree traits, as well as wind conditions and surrounding vegetation. This study of a neotropical canopy tree, Platypodium elegans, explored the extent to which parental variation in diaspore and tree traits explained (1) rate of diaspore descent in still air, (2) distributions of diaspores dispersed from a 40‐m tower in the forest, and (3) natural diaspore distributions around the parent tree. The geometric mean rate of descent in still air among 20 parents was highly correlated with geometric mean wing loading1/2 (r = 0.84). However, diaspore traits and rate of descent predicted less variation in dispersal distance from the tower, although descent rate−1 consistently correlated with dispersal distance. Measured seed shadows, particularly their distribution edges, differed significantly among six parents (DBH range 62–181 cm) and were best fit by six separate anisotropic dispersal kernels and surveyed fecundities. Measured rate of descent and tree traits, combined in a mechanistic seed dispersal model, did not significantly explain variation among parents in natural seed dispersal distances, perhaps due to the limited power to detect effects with only six trees. Seedling and sapling distributions were at a greater mean distance from the parents than seed distributions; saplings were heavily concentrated at far distances. Variation among parents in the distribution tails so critical for recruitment could not be explained by measured diaspore or tree traits with this sample size, and may be determined more by wind patterns and the timing of abscission in relation to wind conditions. Studies of wind dispersal need to devote greater field efforts at recording the “rare” dispersal events that contribute to far dispersal distances, following their consequences, and in understanding the mechanisms that generate them.

No abstract