Eight balloon-borne in situ measurements of CIO in the stratosphere are presented. Six of the observations, which were carried out close to local noon, are compared with recent model calculations at the appropriate latitude (32¿N), altitude (25--45 km), and solar zenith angle interval (10¿<&khgr;<55¿). The CIO mixing ratio by volume observed on four flights between the autumnal equinox and the winter solstice (September 20, October 25, December 2, and December 8, 1977) is found to exhibit little variability, defining an envelope which decreases from 0.5 to 1.1 ppb at 42 km to 10 to 50 ppt at 25 km. Two observations in July (July 28, 1976, and July 14, 1977) indicate significantly higher mixing ratios, the former increasing from 0.8 ppb at 42 km to 1.8 ppb at 36 km an then decreasing to 0.1 ppb at 24 km in reasonable agreement with calculations using 2.2 ppb total chlorine and 8 ppm H2O at the stratopause. The latter summer observation exhibits a serious divergence from calculated densities about 32 km, reaching 7 ppb in a layer between 35 and 40 km. Emphasis is given to a description of laboratory simulation experiments done in parallel with the stratospheric observations and to techniques used for determining the absolute sensitivity of the instrument. Results of the in situ observations are compared with recent results from the laser heterodyne and microwave limb scanning remote sensing methods. The simultaneous observations of and 3> on July 14, 1977, are used in conjunction with a formulation of odd oxygen chemistry in terms of rate limiting steps to deduce that either (a) the or 3> observation is wrong, (b) th observed chlorine was injected into the stratosphere in a period short in comparison with the chemical response time of ozone, or (c) there are important mechanistic steps missing from stratospheric chlorine-ozone photochemistry.