The proctoscopic diagnosis of haemorrhoids may be influenced by the surgeon's knowledge of the presence or absence of associated symptoms. In this study, an observer with no knowledge of the history, was used to check the surgeon's proctoscopic findings in 12 asymptomatic controls, and 24 symptomatic patients on 2 occasions, the latter group undergoing McGivney rubber band ligation. There was very good correlation between the findings of the surgeon and the observer, indicating a lack of 'historical bias'. The documentation method designed to allow this comparison proved sufficiently accurate and reproducible to enable a correlation between haemorrhoidal mass and symptoms. Relief of symptoms after treatment correlated well with an objective reduction in haemorrhoidal mass.