We use in situ measurements of CO obtained in the tropics from 1995 to 1997 on the NASA ER-2 aircraft and a simple photochemical model to calculate the elapsed time between the entry of air into the stratosphere and the observation, which we define as the photochemical age of the air. Assuming this age represents the transit time of the air mass from a boundary at 390 K to its measured altitude, we calculate boundary condition values of CO2 derived from in situ measurements of this species from 400 to 480 K. We validate the approach by comparing these CO2 boundary values with an independent representation of the boundary condition from observations of CO2 in air that had recently entered the stratosphere (as indicated by simultaneous measurements of N2O, CO, and potential temperature). For five of the six flights, differences between CO2 boundary condition values determined using the photochemical age of the air and those derived from independent measurements can be accounted for with isentropic mixing of midlatitude stratospheric air into the tropics. Having validated the photochemical ages of the sampled stratospheric air, we use the same analysis of in situ water data to provide water vapor boundary condition values that constrain the seasonal cycle of water vapor entering the stratosphere. On the basis of these constraints we evaluate the seasonal cycle of entry-level water vapor derived from tropical tropopause temperatures from the radiosonde network between 10 ¿S and 10 ¿N. We conclude that while average saturation mixing ratios provide a suitable boundary condition for water vapor entering the stratosphere, the uncertainties in saturation mixing ratios derived from radiosonde temperatures and the lack of coverage prevents distinguishing between ascent preferentially occurring over the western equatorial Pacific or throughout the tropics. With the assumption that vertical diffusion and midlatitude mixing have a negligible effect on the calculated age, ascent velocities can be inferred from the photochemical ages. These ascent velocities show a seasonal cycle that is inconsistent with our current understanding of the dynamics driving the stratospheric circulation and with independent estimates of tropical vertical ascent rates. ¿ 2001 American Geophysical Union