Although exchange of genetic information by recombination plays a role in the evolution of viruses, the extent to which it generates diversity is not clear. We analyzed genomes of geminiviruses for recombination using a new statistical procedure developed to detect gene conversions. Geminiviruses (family, Geminiviridae) are a group of plant viruses characterized by a genome of circular single-stranded DNA (approximately 2700 nucleotides in length) encapsidated in twinned quasi-isometric particles. Complete nucleotide sequences of geminiviruses were aligned, and recombination events were detected by searching pairs of viruses for sequences that are significantly more similar than expected based on random distribution of polymorphic sites. The analyses revealed that recombination is very frequent and occurs between species and within and across genera. Tests identified 420 statistically significant recombinant fragments distributed across the genome. The results suggest that recombination is a significant contributor to geminivirus evolution. The high rate of recombination may be contributing to the recent emergence of new geminivirus diseases.

We present a worldwide analysis of humid tropical forest dynamics and tree species richness. New tree mortality, recruitment, and species richness data include the most dynamic and diverse mature tropical forests known. Twenty-five sites show a strong tendency for the most speciesrich forests to be dynamic and aseasonal. Mean annual tree mortality and recruitment-turnover-is the most predictive factor ofspecies richness, implying that small-scale disturbance helps regulate tropical forest diversity. Turnover rates are also closely related to the amount of basal area turnover in mature tropical forests. Therefore the contribution of small-scale disturbance to maintaing tropical forest diversity may ultimately be driven by ecosystem productivity.The inverse relationship of species richness with latitude has long attracted ecologists' interest. Some explanations concentrate on the evolutionary origins ofhigh tropical diversity; others focus on its maintenance. High community and regional diversity in the tropics has been linked to available energy (1-3), species-specific predation (4), and delayed competitive exclusion (5-7). Yet, the existence of major differences in species richness among tropical forests, of similar theoretical interest, has only recently been well documented (8)(9)(10)(11)(12).Several hypotheses link tree diversity to disturbance at varying spatial and temporal scales. Frequency-or densitydependent mortality may enhance diversity (13). Locally unpredictable disturbance may counteract the effects of pairwise, predictable competitive displacement (7). Environmental contrasts within and among tree-fall gaps may favor trees with different regeneration requirements (14,15 (17) demonstrated density-dependent mortality affecting the most abundant tree species in Barro Colorado Island, Panama. Yet, few coincident data on tropical forest species richness and turnover have been reported, hindering biome-wide evaluation of the dynamics-diversity relationship. In this paper we show that variation in tree species richness among tropical forests not subject to largescale natural disturbances is related to their dynamism, defined as the mean rate of mortality and recruitment, which in turn is related to productivity. We hypothesize that productivity ultimately drives long-term ecosystem dynamics in mature tropical forests, since the faster a forest grows the more rapid stem turnover must be, as long as total biomass is capped by constraining factors. Therefore, high productivity may help maintain exceptional species richness through promoting frequent, spatially unpredictable small-scale disturbance. METHODSWe investigate dynamics and species richness at 25 mature forests from all major tropical regions. Original data come from Peru and Sarawak, where trees -10 cm in diameter at breast height (d.b.h.) were identified, with additional data from the literature and colleagues

Population geneticists have long sought to estimate the distribution of selection intensities among genes of diverse function across the genome. Only recently have DNA sequencing and analytical techniques converged to make this possible. Important advances have come from comparing genetic variation within species (polymorphism) with fixed differences between species (divergence). These approaches have been used to examine individual genes for evidence of selection. Here we use the fact that the time since species divergence allows combination of data across genes. In a comparison of amino-acid replacements among species of the mustard weed Arabidopsis with those among species of the fruitfly Drosophila, we find evidence for predominantly beneficial gene substitutions in Drosophila but predominantly detrimental substitutions in Arabidopsis. We attribute this difference to the Arabidopsis mating system of partial self-fertilization, which corroborates a prediction of population genetics theory that species with a high frequency of inbreeding are less efficient in eliminating deleterious mutations owing to their reduced effective population size.

Sexual dimorphism is widespread and substantial throughout the animal world. It is surprising, then, that such a pervasive source of biological diversity has not been integrated into studies of adaptive radiation, despite extensive and growing attention to both phenomena. Rather, most studies of adaptive radiation either group individuals without regard to sex or focus solely on one sex. Here we show that sexual differences contribute substantially to the ecomorphological diversity produced by the adaptive radiations of West Indian Anolis lizards: within anole species, males and females occupy mostly non-overlapping parts of morphological space; the overall extent of sexual variation is large relative to interspecific variation; and the degree of variation depends on ecological type. Thus, when sexual dimorphism in ecologically relevant traits is substantial, ignoring its contribution may significantly underestimate the adaptive component of evolutionary radiation. Conversely, if sexual dimorphism and interspecific divergence are alternative means of ecological diversification, then the degree of sexual dimorphism may be negatively related to the extent of adaptive radiation.

Statistical tests for detecting gene conversion are described for a sample of homologous DNA sequences. The tests are based on imbalances in the distribution of segments on which some pair of sequences agrees. The methods automatically control for variable mutation rates along the genome and do not depend on a priori choices of potentially monophyletic subsets of the sample. The tests show strong evidence for multiple intragenic conversion events at two loci in Escherichia coli. The gnd locus in E. coli shows a highly significant excess of maximal segments of length 70-200 bp, which suggests conversion events of that size. The data also indicate that the rate of these short conversion events might be of the order of neutral mutation rate. There is also evidence for correlated mutation in adjacent codon positions. The same tests applied to a locus in an RNA virus were negative.