The detection of N2O5 in the stratosphere was reported by Toon et al. (1986) and Toon (1987) based on measurements of the absorption by the N2O5 bands at 1246 and 1720 cm-1 in solar occultation spectra recorded at sunrise near 47¿S latitude by the Atmospheric Trace Molecule Spectroscopy (ATMOS) experiment during the Spacelab 3 (SL3) shuttle mission (April 29 to May 6, 1985). In this work we report additional measurements and analysis of stratospheric N2O5 derived from the ATMOS/SL3 spectra. The primary results are the detection and measurement of N2O5 absorption at sunset in the lower stratosphere, the inversion of a precise (~10%) N2O5 sunrise vertical distribution between 25.5- and 37.5-km altitude, and the identificateion and measurement of absorption by the N2O5 743-cm-1 band at sunrise. Assuming 4.32¿10-17 and 4.36¿10-17 cm-1/molecule cm-2 for the integrated intensities of the 1246- and 743-cm-1 bands, respectively, at stratospheric temperatures, retrieved volume mixing ratios (VMRs) in parts per billion by volume (ppbv) at sunrise (47¿S latitude) are 1.32¿0.34 at 37.5 km, 1.53¿0.35 at 35.5 km, 1.63¿0.36 at 33.5 km 1.60¿0.34 at 31.5 km, 1.43¿0.30 at 29.5 km, 1.15¿0.24 at 27.5 km, nd 0.73¿0.15 at 25.5 km. Retrieved VMRs in ppbv at sunset (31¿N latitude) are 0.13¿0.05 at 29.5 km, 0.14¿0.05 at 27.5 km, and 0.10¿0.04 at 25.5 km. Quoted error limits (1 sigma) include the error in the assumed band intensities (~20%). Within the error limits of the measurements the inferred mixing ratios at sunrise agree with diurnal photochemical model predictions obtained by two groups using current photochemical data. The measured mixing ratios at sunset are lower than the model predictions with differences of about a factor of 2 at 25- km altitude. ¿ American Geophysical Union 1989