Interacting components of interspecific relative growth rate: constancy and change under differing conditions of light and nutrient supply

1. One thousand three hundred and twenty plants from 22 species of herbaceous angiosperms, typical of open sunny habitats, were grown from seed for 35 days under controlled conditions in four experimental environments: high (1100 mu mol m(-2) s(-1) PAR) and low (200 mu mol m(-2) s(-1) PAR) light, combined with high (full-strength) and low (1/6 dilution) levels of hydroponic nutrient solution. Plants of each species were harvested at 15, 20, 25, 30 and 35 days post-germination. 2. Relative growth rate (RGR), net assimilation rate (NAR), specific leaf area (SLA) and leaf mass ratio (LWR) were estimated for each species in each treatment combination. 3. Both light and nutrient levels affected RGR and each of the growth components. Average NAR increased, while SLA and LWR decreased, with increasing light. Increasing nutrient levels increased all three growth components. Light and nutrient levels interacted in their effects on NAR and SLA, but not in their effects on LWR. 4, Those species having the highest RGR in the most productive environment were most severely depressed in the less productive environments. This same pattern also occurred with NAR and SLA, but not with LWR. 5. The rank order of species with respect to SLA and LWR remained similar across the treatment environments; with respect to NAR the rank order remained similar in only four of the six treatment contrasts; with respect to RGR the rank order remained similar in only two of the six treatment contrasts and the rank order actually reversed in one treatment contrast. 6. RGR did not display strong bivariate correlations with any of the growth components in any of the treatment groups. It had positive, moderately strong correlations with NAR but only in the low-nutrient treatments. It had positive, moderately strong correlations with LWR only in the high-nutrient treatments. After controlling for the other growth components, each showed strong positive relationships with RGR. Furthermore, the effects of the treatments disappeared after controlling for the components, showing that the treatment effects were completely mediated through their effects on these underlying components. 7. There were also strong negative partial correlations between each of the growth components. These compensatory relationships buffered RGR from large changes in response to changing environments.