A large number of ozone profiles measured by using various methods (lidar, ozonesondes, microwave radiometer, and Stratospheric Aerosol and Gas Experiment II) over Lauder, New Zealand, between late 1994 and early 1998 are intercompared. These profiles are also used to validate a collocated Dobson spectrophotometer. Between March 1996 (August 1996 for sonde measurements) and January 1998, all instruments were operational and no instrument changes took place. The ozone number densities averaged over this period agreed within 5% in the 20- to 35-km range. Between 12 and 20 km, lidar and sonde results deviated by less than 8%, and lidar and SAGE II results deviated by less than 15%. The ozone column densities measured by the Dobson spectrophotometer agreed within 3% with the integrated ozone profiles. The various methods are discussed, and modifications are proposed. They comprise a decrease of the reported sonde altitudes of about 125 m (¿50 m) to correct for the response time of the sonde's chemical solution, the use of more accurate molecular parameter values in the lidar algorithm, and a 2.5% decrease in SAGE II ozone densities. Improved agreement between the average ozone profiles to within 1.5% for lidar and sondes (20--35 km) and within 2.5% for lidar and SAGE II (20--35 km) is achieved. The Dobson results are found to be influenced by the annual cycle of the temperature profile through the altitude-averaged ozone absorption cross section, which has been assumed to be constant in the presented data but actually varies with an amplitude of 2% over Lauder. ¿ 2000 American Geophysical Union