In this review of English language publications from 1970, 5-year survival rates after surgery for gastric cancer have been analysed. While the proportion of patients coming to operation has fallen from 92 per cent before 1970 to 71 per cent by 1990, the proportion of operated patients undergoing resection has increased from 37 per cent before 1970 to 48 per cent before 1990. This change suggests improved preoperative staging leading to better patient selection for operation. The 5-year survival rate following all resections has increased significantly from 20.7 per cent before 1970 to 28.4 per cent before 1990, an increase of 7.7 per cent (95 per cent confidence interval 7.1–8.3 per cent). The 5-year survival rate following curative or radical resection has risen from 37.6 to 55.4 per cent over the same period, an increase of 17.8 per cent (95 per cent confidence interval 17.1–18.5 per cent). It is likely that this improvement has contributed to the decrease in the mortality rate from gastric cancer. Comparison of Japanese series with others suggests that diagnosis and treatment of the disease at an earlier stage will result in an even greater increase in 5-year survival rates outside Japan. Of the papers studied, 56 per cent were excluded from analysis, the majority because the data provided about 5-year survival rates were insufficient or the survival calculations inappropriate. Results of survival after operations for gastric cancer should be calculated and presented in a standardized manner.