The top edge (20--22 km) of the Antarctic ozone hole has been identified as a good place to detect ozone recovery (D. J. Hofmann et al., 1997). During the first six POAM observation years (1994--1996, 1998--2000), the early October ozone mixing ratio at the top edge of the ozone hole was nearly constant, while during the last four years (2001--2004) it was noticeably greater. The estimated ozone photochemical loss for air sampled at 20--22 km in 2001--2004 was likewise smaller. During 2001--2004, there were fewer PSCs and generally higher temperatures during August and September at 20--22 km than prior years. The increased ozone was due to both reduced photochemical ozone loss (2001, 2003, and 2004) and enhanced adiabatic descent of the ozone profile (2002). Because of the changing meteorological conditions, the ozone changes at the top edge of the hole cannot be easily attributed to changes in chlorine abundance.