A SAOZ (System d’Analyse et d’Observations Zenithales) u.v.-visible spectrometer was installed at Faraday, Antarctica, in 1990, with a view to gaining year-round measurements of column ozone. Observations have been validated by comparison with the Faraday Dobson. Agreement between the two instruments is good, with SAOZ measuring 3 ± 2% lower than the Dobson over the comparison period. Both instruments show the same small-scale variation in column ozone. However, there is a notable seasonal signal in the ratio of (SAOZ column ozone)/(Dobson column ozone) which calls into question the validity of the SAOZ winter column ozone measurements made when the solar elevation is too low for making Dobson observations. The problem appears to lie with the calculated air-mass factors (AMFs) used in deriving vertical columns from the SAOZ slant columns. New SAOZ data were derived using empirical AMFs. The qualitative behaviour of winter column ozone is the same for both sets of AMFs, showing an increase to mid-winter, followed by a decrease to the spring minimum, although smaller mid-winter columns are calculated when using empirical AMFs.