In an effort to better constrain atmospheric water vapor mixing ratios and to understand the discrepancies between different measurements of water vapor in the stratosphere and troposphere, we have carefully examined data from the Harvard Lyman-α photofragment fluorescence hygrometer, which has flown on the NASA ER-2 aircraft from 1992 through 1998. The instrument is calibrated in the laboratory before and after each deployment, and the calibration is checked by direct absorption measurements in the troposphere. On certain flights, the ER-2 flew level tracks during which water vapor varied by up to 80 ppmv, under nearly constant atmospheric conditions. These flights provide a stringent test of our calibration via direct absorption and indicate agreement to within 3%. During the 1997 Photochemistry of Ozone Loss in the Arctic Region In Summer (POLARIS) mission, our Lyman-α instrument was compared with a new diode laser hygrometer from the Jet Propulsion Laboratory. Overall agreement was 5% during the June/July deployment and 1% for potential temperatures of 490 to 540 K. The accuracy of our instrument is shown to be ¿5%, with an additional offset of at most 0.1 ppmv. Data from this instrument, combined with simultaneous measurements of CH4 and H2, are therefore ideal for studies of the hydrogen budget of the lower stratosphere. ¿ 1999 American Geophysical Union