The isotopic composition of long-lived trace molecules provides a window into atmospheric transport and chemistry. Carbon dioxide is a particularly powerful tracer, because its abundance remains >100 parts per million by volume (ppmv) in the mesosphere. Here, we successfully reproduce the isotopic composition of CO 2 in the middle atmosphere, which has not been previously reported. The mass-independent fractionation of oxygen in CO 2 can be satisfactorily explained by the exchange reaction with O( 1 D). In the stratosphere, the major source of O( 1 D) is O3 photolysis. Higher in the mesosphere, we discover that the photolysis of 16 biogeochemical cycles ͉ CO2 ͉ mesosphere ͉ stratosphere O f the many trace molecules that can be used to examine atmospheric transport processes and chemistry [e.g., CH 4 , N 2 O, SF 6 , and the chlorofluorocarbons (CFCs)], carbon dioxide is unique in the middle atmosphere, because of its high abundance [Ϸ370 parts per million by volume (ppmv) in the stratosphere, dropping to Ϸ100 ppmv at the homopause]. The massindependent isotopic fractionation of oxygen first discovered in ozone (1-5) is thought to be partially transferred to carbon dioxide (3, 6-11) via the reaction O( 1 D) ϩ CO 2 in the middle atmosphere (12, 13). Indeed, whereas the reactions of trace molecules with O( 1 D) usually lead to their destruction (14) ) than those seen at 30°N over a similar altitude range and a slope of m ϭ 1.72 Ϯ 0.22 (2 ). Over smaller altitude intervals, a similar analysis yields m ϭ 2.06 Ϯ 1.16 (2 ) for samples from the Arctic vortex (6) and 1.64 Ϯ 0.38 (2 ) from the lower stratosphere (7). Slopes near 1.6-1.7 have been successfully reproduced in laboratory photochemical experiments under approximately stratospheric conditions (17).Despite the large uncertainties in m, induced by the challenges associated with the sample collection and mass spectrometry of middle atmospheric O 3 (2, 5) and CO 2 (8, 10, 11), it is worth examining the potential sources of variation in the isotopic composition of carbon dioxide, both to better understand the atmosphere and as a guide for future observations. Yung et al.(13), for example, have suggested that the upwelling of tropospheric air from the tropics along with downwelling near 30°N could dilute the magnitude of the fractionation induced by photochemistry. It is beyond the scope of this work to provide a detailed explanation of the effect of transport on the isotopic composition of CO 2 . Briefly, the key concept is the ''age of air,'' with the clock being set to zero at the tropopause. As the air parcel travels to the middle atmosphere, it ages, and there is more time for photochemical equilibrium to be reached. Morgan et al. (18) discuss these issues at length and give a heuristic relation between the intrinsic enrichment factor ( ) and the resulting isotopic composition (␦) for the one-dimensional (1D) diffusion-limited transport case:where r ϭ chem /4 trans with chem ϭ chemical lifetime (here the isotopic exchange time) and trans ϭ transport time (here...