What was done
Eight beech (Fagus sylvatica L.) and eight spruce (Picea abies Karst.) seedlings, along with five to six individuals from each of several typical understory species, were planted in each of 32 lysimeters that were enclosed in pairs in 16 open-top chambers (OTCs) in the field at Birmensdorf, Switzerland.  One lysimeter in each OTC was filled with a nutrient-poor acidic soil, while the other lysimeter was filled with a fertile calcareous soil.  Adding yet another variable to the mix, the eight beech seedlings were obtained from four different provenances (populations), while the eight spruce seedlings were also selected from eight different provenances, but in this case from two populations and six clones.  Under these conditions, the trees were grown for four full years, half of them in ambient air having a mean CO2 concentration of 370 ppm, and half of them in CO2-enriched air having a mean concentration of 570 ppm.

What was learned
Total biomass growth responses to the 200-ppm increase in the air's CO2 concentration ranged from -26% to -7% across the beech provenances, but from +9% to +38% across the spruce provenances, when the trees were rooted in the nutrient-poor acidic soil.  In contrast, when the trees were rooted in the fertile calcareous soil, the growth responses ranged from -4% to +40% for beech and from +10% to +74% for spruce.  Also of note was the fact that in this experiment, where the trees competed inter- and intra-specifically for light, space and nutrients, the most CO2-responsive provenances when grown on the acidic soil were sometimes not the most responsive provenances when grown on the calcareous soil.

What it means
The authors say their study "demonstrates that the soil is a major determinant of plant responses to environmental changes, and that intraspecific differences in growth-sensitivity to CO2 ? are considerable, especially on fertile soils," adding that "rather different conclusions would have been reached had we chosen to work with only one type of soil, or with a single provenance of each species."  Hence, they conclude that this finding "underlines the need for rather complex experimental and statistical designs in order to reach a better understanding of forest tree responses to environmental changes."