Total cigarettes (all brands) sold weekly by a panel of 60 New Zealand supermarkets were monitored electronically for 42 weeks, a period when cigarette advertisements were in plain format with strong, varied disease warnings. Real cigarette price, newspaper advertising of old, regular and upmarket brands, and the number of newspaper news items on smoking issues were inversely associated with cigarette sales. Tending to increase total sales (all brands) were: more non-shopping days in the current week, and in the week following; volume of grocery items purchased, to indicate income and store traffic; and real advertising expenditure in newspapers for new downmarket cigarette brands, particularly one heavily-advertised brand (Peter Jackson) which was in late 1989 smoked by 4% of teenage smokers. All factors when interacting, explained 93% of changes in weekly cigarette sales. Most of the change occurred in the same week, and was 90% in place after a further 3 weeks. Newspapers, by doubling news coverage of smoking issues or by banning cigarette advertisements, can lower cigarette consumption as much as can a 10% price increase.