The increase in number of fine (< 0.5 mm diameter) roots of one-year-old clonal Pinus radiata D. Don trees grown in large open-top field chambers at ambient (362 micro mol mol(-1)) or elevated (654 micro mol mol(-1)) CO(2) concentration was estimated using minirhizotron tubes placed horizontally at a depth of 0.3 m. The trees were well supplied with water and nutrients. Destructive harvesting of roots along an additional tube showed that there was a linear relationship between root number estimated from the minirhizotron and both root length density, L(v), and root carbon density, C(v), in the surrounding soil. Root distribution decreased with horizontal distance from the tree. At a depth of 0.3 m, 88% of the total C(v) was concentrated within a 0.15-m radius from tree stems in the elevated CO(2) treatment, compared with 35% for trees in the ambient CO(2) treatment. Mean C(v) along the tubes ranged up to 5 x 10(-2) micro g mm(-3) and tended to be greater for trees grown at elevated CO(2) concentration, although the differences between CO(2) treatments were not significant. Root growth started in spring and continued until late summer. There was no significant difference in seasonal rates of increase in C(v) between treatments, but roots were observed four weeks earlier in the elevated CO(2) treatment. No root turnover occurred at a depth of 0.3 m during the first year after planting. Mean values of carbon dioxide flux density at the soil surface, F, increased from 0.02 to 0.13 g m(-2) h(-1) during the year, and F was 30% greater for trees grown at elevated CO(2) concentration than at ambient CO(2). Diurnal changes in F were related to air temperature. The seasonal increase in F continued through the summer and early autumn, well after air temperature had begun to decline, suggesting that the increase was partly caused by increase in C(v) as the roots colonized the soil profile.