The ‘social control’ and ‘incongruence’ hypotheses, first put forward by ROHWER (1977) to explain how attempted ‘deceit’ status signalling is kept in check among winter-flocking birds, were tested under semi-natural conditions for Parus major. This species signals its social status by the width of its breast stripe. The lowest-ranked male in experimental flocks, each made up of four individuals, was manipulated in one of three ways: 1) the status signal was altered by painting the breast stripe to make it broader; 2) agonistic behaviour was altered by injecting testosterone; 3) both status signal and behaviour were manipulated. A study of the outcome of subsequent agonistic encounters by these birds revealed that the status of the manipulated individuals only rose when both their behaviour and status signal were altered. This indicates that the ‘social control’ hypothesis must be rejected, but not the ‘incongruence’ hypothesis.