Diurnal variation of peak expiratory flow was assessed over 10 days in 40 asthmatic outpatients judged clinically to be stable. Thirty patients showed a morning dip in their peak expiratory flow and the frequency with which this occurred appeared to be linked with the amplitude of the dip. The percentages showing large, moderate and insignificant dips were similar to those described in asthmatics hospitalised with an exacerbation. Ten patients were admitted with an exacerbation of their asthma in the preceding 2 years. Two of the three non-dippers assumed a dipping pattern during this admission. Without a continuous peak expiratory flow record it would not have been thought that many of our apparently stable patients were showing high amplitude dips such as have been associated with sudden unexpected death. The place of a continuous diurnal peak expiratory flow record in relation to asthma mortality is discussed.