The stable C isotope composition (d 13 C) of CO 2 respired by trunks was examined in a mature temperate deciduous oak forest (Quercus petraea). Month-to-month, day-to-day and diurnal, measurements were made to determine the range of variations at different temporal scales. Trunk growth and respiration rates were assessed. Phloem tissue was sampled and was analysed for total organic matter and soluble sugar 13 C composition. The CO 2 respired by trunk was always enriched in 13 C relative to the total organic matter, sometimes by as much as 5&. The d 13 C of respired CO 2 exhibited a large seasonal variation (3.3&), with a relative maximum at the beginning of the growth period. The lowest values occurred in summer when the respiration rates were maximal. After the cessation of radial trunk growth, the respired CO 2 d 13 C values showed a progressive increase, which was linked to a parallel increase in soluble sugar content in the phloem tissue (R = 0.95; P < 0.01). At the same time, the respiration rates declined. This limited use of the substrate pool might allow the discrimination during respiration to be more strongly expressed. The lateseason increase in CO 2 d 13 C might also be linked to a shift from recently assimilated C to reserves. At the seasonal scale, CO 2 d 13 C was negatively correlated with air temperature (R = -0.80; P < 0.01). The diurnal variation sometimes reached 3&, but the range and the pattern depended on the period within the growing season. Contrary to expectations, diurnal variations were maximal in winter and spring when the leaves were missing or not totally functional. By contrast to the seasonal scale, these diurnal variations were not related to air temperature or sugar content. Our study shows that seasonal and diurnal variations of respired 13 C exhibited a similar large range but were probably explained by different mechanisms.

In situ (13)C/(12)C isotopic labelling was conducted in field-grown beech (Fagus sylvatica) twigs to study carbon respiration and allocation. This was achieved with a portable gas-exchange open system coupled to an external chamber. This method allowed us to subject leafy twigs to CO(2) with a constant carbon isotope composition (delta(13)C of -51.2 per thousand) in an open system in the field. The labelling was done during the whole light period at two different dates (in June 2002 and October 2003). The delta(13)C values of respiratory metabolites and CO(2) that is subsequently respired during the night were measured. It was found that night-respired CO(2) is not completely labelled (only ca. 58% and 27% of new carbon is found in respired CO(2) immediately after the labelling in June 2002 and October 2003, respectively) and the labelling level progressively disappeared during the next day. It is concluded that the carbon respired by beech leaves after illumination was supplied by a mixture of carbon sources in which current carbohydrates were not the only contributors. In addition, as has been found in herbaceous plants, isotopic data before labelling showed that carbon isotope discrimination favoring the (13)C isotope occurred during the night respiration of beech leaves.