Frequency distributions compiled from data in two unrelated judgement situations were analysed with regard to the greater frequency of certain numbers relative to other numbers. In a laboratory situation subjects were requested to give their ‘best estimates' of the distance between two tachistoscopically presented points of light. In the second situation, data were collected from hunters who estimated the number of quail they had killed during the preceding season. The frequency distributions from both studies contained significantly more anchored than non-anchored judgements. Analysis of the numbers given in the laboratory (hex situation) study also showed a strong preference for ‘multi-anchored’ rather than non-anchored numbers. In the field (quail-kill) study the ‘triple-anchored’ numbers were the estimates given with proportionately the greatest frequency.