In situ measurements of water vapor in the stratosphere with a new instrument are reported.The instrument has been designed to observe daytime water vapor from a multi-instrument balloon gondola that simultaneously measures free radicals such as OH, HOz, and Oa in the stratosphere up to 40 km. Lyman-alpha photofragment fluorescence is used to measure water molecules in a flowing sample of ambient air. Outgassing from the interior walls of the instrument is avoided by cooling the walls with liquid nitrogen to a temperature near or below the dewpoint of the environment and by drawing air through the instrument with a fan. A brief description of the instrument is given, followed by the results of the first four balloon flights. Because frost formation in the scattering chamber resulted in a large and variable background, the data from July 15, 1987, have a relatively modest signal-to-noise ratio. The measured mixing ratio for this flight varies from 3.0-5.5 ppmv over the altitude range of 17-34 kin. Adjustments in the cooling protocol for the flights of July 6, 1988, July 28, and August 25, 1989, result in a much higher signal-to-noise ratio. Profiles from these three flights are similar to, but somewhat higher, than the 1987' profile. The July 1988 and August 1989 profiles exhibit the highest mixing ratios, reaching peak values of about 6.5 ppmv near 35 kin. Implications of these four measurements are discussed, as are the issues of short-and long-term variability of stratospheric water vapor. to a secular trend of increasing water in the stratosphere [Ehhalt, 1986; Blake and Rowland, 1988]. The role of water vapor in the stratosphere takes on added significance in light of the dramatic ozone depletion over Antarctica and the severely perturbed chemical environment of the polar vortex regions [Anderson et al., 1989; Brune et al., 1990; Salawitch et al., 1990]. In particular, the formation of polar stratospherlc clouds depends upon the water vapor mixing ratio as well as the temperature. Water vapor concentration and trends should be reflected in the frequency, extent, and duration of PSCs and, by extension, in the perturbed nitrogen and chlorine chemistry observed in the polar stratosphere during the winter-spring transition. The generally accepted morphology of stratospheric water is based on the LIMS measurements from the NIMBUS 7 satellite [Russell et al., 1984; Remsberg et al., 1984], but typically little regard is given to the trend in water when these measurements are cited and used. While the LIMS H20 data have been subject to careful analysis and validated by a number of in situ profile measurements, it is extremely difficult to assess impartially the accuracy of satellite profile measurements. In the case of ozone measurements, there has been an ongoing systematic difference between in situ and satellite profile measurements [Hilsenrath et al., 1986; Robbins, 1987]. In particular, it should be stressed that satellite measurements are remote measurements, and yield water vapor mixing ratios only after exten...