That populations can adapt their safer-sex behaviour to continue to protect themselves when their habits change is evidenced by the Wawer study quoted above (Wawer 2005). In her study of adults in Rakai, Uganda Wawer found that although HIV prevalence had declined by 75 per cent between 1993 and 2004, HIV incidence had not, and was running at a steady annual seroconversion rate of 1.5 per cent or so. The decline in the number of people living with HIV in the area was almost entirely due to the thinning of the population by AIDS.

She found evidence that people were actually having more extramarital sex and having it younger in 2004 than in 1993, and that the behavioural changes of the early 1990s were being reversed. For instance, the proportion of 15 to 19 year olds who were sexually active had gone up from 40 to 50 per cent, and the number of adult men reporting two or more partners a year had increased from 20 to 27 per cent.

What was keeping the incidence rate steady in the face of this ‘liberalising’ of sexual behaviour was an increased level of condom use. Condom use among men in general with casual partners had increased from 10 per cent in 1993 to 50 per cent in 2004 – a figure described by Wawer as ‘incredibly high by African standards’ (see columns J and K above).

This figure of 50 per cent was skewed by a 95 per cent rate of condom use in the few per cent of men who admitted having commercial sex. But even in male teenagers, who by and large did not use sex workers, it had gone up from 19 to 38 per cent.

Other recent surveys have reported figures of about 50 per cent of men in South Africa (Peltzer 2000) and Uganda (Najjumba 2003) saying they had ‘ever’ used condoms, with considerably higher usage in sex perceived to be risky.