We studied the effects of CO 2 elevation on leaf and canopy photosynthesis and optimal leaf area index (LAI) for stands of the annual species Abutilon theophrasti and Ambrosia artemisiifolia. Leaf photosynthesis was modeled as a function of photosynthetic photon flux density (PPFD) and nitrogen content per unit leaf area (N L ). There was a curvilinear relationship between the light-saturated rates of leaf photosynthesis (P max ) and N L . CO 2 elevation significantly increased P max as a function of N L in both species. Dark respiration (R d ) was linearly correlated with N L . CO 2 elevation slightly but significantly increased R d in Abutilon, while it had no significant effect on R d in Ambrosia. The initial slope of a light-response curve was determined from quantum yield ( abs ) multiplied by leaf absorptance and then calibrated against N L . Daily canopy photosynthesis, calculated by integration of leaf photosynthesis with the actual distribution of leaf area, leaf N, and PPFD within a canopy, showed fairly good agreement with the canopy photosynthesis estimated from growth analysis. CO 2 elevation increased canopy photosynthesis by 30-50%. Based on the leaf photosynthesis model for Abutilon, we calculated daily canopy photosynthesis for a given LAI and N availability, in which N was assumed to be distributed optimally within a leaf canopy to maximize daily canopy photosynthesis. An optimal LAI to maximize daily canopy photosynthesis was obtained for each level of N availability and this optimum increased with increasing N availability. Contrary to the often predicted increase in LAI with CO 2 elevation, the optimum LAI did not increase at high CO 2 when N availability was limited. Two factors were suggested to be involved in counteracting the increase in LAI in a high-CO 2 world. One is the higher abs of plants grown in elevated CO 2 , which makes leaves in the canopy more N limited, favors higher N L and thus lowers optimal LAI. The other is the higher R d in elevated CO 2 , which leads to higher light compensation points, and lowers optimal LAI.

Summary 1The effect of elevated CO 2 on species' performance was investigated in communities composed of five annual weeds that are characteristic of early old field succession in central Europe: Centaurea cyanus L., Matricaria chamomilla L., Silene noctiflora L., Papaver rhoeas L. and Legousia speculum-veneris (L.) Chaix. 2 The experiment was based on a simplex design, repeated at two overall levels of initial stand density, to give a wide range of five-species communities across which the initial composition and species abundance varied systematically. 3 A multivariate method, based on analysing the differences in relative growth rates (RGRD) between pairs of species, was extended for use with more than two species, in order to assess the relative importance of various determinants of change in stand biomass composition. 4 On average, Centaurea (54.6% of final yield) gave the highest yield, followed by Matricaria (22.9%), Silene ( 16.9%), Legousia (3.1%) and Papaver (2.7%). 5 The major determinants of change in community structure were species identity and CO 2 level. Elevated CO 2 significantly changed community composition towards the previously more poorly performing species Silene , Legousia and Papaver. 6 Despite strong effects of intra-and interspecific competition on individual species performance, species' initial abundance had relatively little impact on the change in community composition. Most cases where such effects were significant involved Silene : performance of Papaver was poorer in communities with higher initial presence of Silene and higher initial abundances of Centaurea and Matricaria always facilitated performance of Silene. 7 These new methods proved a powerful system for identifying the biotic and abiotic factors determining change in biomass composition in multispecies communities.

Summary1 Neighbour size and distance are confounded in most studies of plant growth and competition at the individual level. To investigate the effects of neighbour size on competitive effect, we grew target Kochia scoparia individuals surrounded by six equidistant, even-aged conspecific neighbours. We varied neighbour size by sowing groups of neighbours at different times, and we also varied the sowing time of the target plants to generate variation in both neighbour and target size during the process of competition. We analysed the growth of target plants over two time intervals as a function of their own size and the size of their neighbours at the beginning of the interval. 2 When competition became intense, the relative growth rate of target plants was primarily determined by the size of their neighbours. There was a negative linear relationship between the relative growth rate of target plants over an interval and the biomass of their neighbours at the beginning of the interval. The size of the target plant itself did not make a significant additional contribution to predicting its growth rate. There is a limit on the growth in biomass of the population (target + neighbours), and growth of individuals occurs within this constraint. Local biomass density, which can be primarily determined by neighbouring individuals, can be much more important for an individual's growth than its own size. 3 There was no evidence of size-asymmetric competition. The size of neighbours was the primary determinant of a target plant's relative growth rate, but the effect of a given amount of neighbour biomass was the same for neighbours larger and those smaller than the target plant.

The ‘Vallée de Mai’ on the island of Praslin in the Seychelles group is a World Heritage Site because of its unique forest vegetation dominated by the palm Lodoicea maldivica. However, despite its great conservation importance, the stand structure, species diversity and regeneration of this forest have not previously been studied. Moreover, the conservation value of the reserve is currently threatened by invading alien plants and limited regeneration of some native species. The species composition of the forest was studied in six permanent plots (1985 and 1998) and along seven trail transects (1997). To explain potential variation in tree regeneration, the light climate within all plots was investigated using hemispherical photographs. The results reveal little change in forest structure, species diversity and regeneration in the permanent plots over the 13‐year period, but some regeneration of most woody species, presence of some alien species and a relatively high species turnover. We found no significant influence of local light climate upon forest structure, species diversity or regeneration. Although the stand characteristics seem to be relatively stable, the permanent plots and transects should be monitored in order to detect future changes in stand structure and to optimize the protection of this unique forest reserve.

It is known that convergence and divergence can occur in complex plant communities, but the relative importance of biotic and abiotic factors driving these processes is less clear. We addressed this issue in an experiment using a range of mixed stands of five species that are common in Swiss fens (Carex elata, C. flava, Lycopus europaeus, Lysimachia vulgaris and Mentha aquatica) and two levels of water and nutrients. One hundred and seventy-six experimental mixtures were maintained in large pots (75 l) for two consecutive growing seasons in an experimental garden. The stands varied systematically in the initial relative abundance of each of the five species and in overall initial stand abundance. The changes in biomass over 2 years were modelled as linear functions of treatments and the initial biomass of each species. The dynamics of the system were mainly driven by differences in the identity of species and by a negative feedback mechanism but also by different abiotic conditions. In all mixtures, C. elata became more dominant over time, which caused an overall convergence of community composition. In addition, the rate of change of each species' biomass was negatively related to its own initial abundance. Thus, a negative feedback further contributed to the convergence of communities. Species responded differently to water level and nutrient supply, causing community dynamics to differ among treatments. However, the different abiotic conditions only slightly modified the overall convergence pattern. Competitive interactions between more than two species were weaker than the negative feedback but still significantly influenced the species' final relative abundance. The negative feedback suggests that there is niche partitioning between the species, which permits their coexistence.