In a paper in the January issue of this volume of Ageing & Society, Eric Midwinter argued that ‘much can be learned from re-drawing the demographic map with social rather than chronological contours’. This opinion reflects a widespread view among social gerontologists that chronological age is an ‘empty’ variable, even though it is central to the construction of social identities, both in bureaucratic contexts and in less formal social interaction. This paper draws on material stored in the Mass-Observation Archive at the University of Sussex, England. A large panel of ‘ordinary people’ was asked to write about ‘growing older’ in 1992 and about ‘birthdays’ in 2002. An analysis of the ways in which they revealed their age demonstrates that the revelation of chronological age is unproblematic in certain contexts that are deemed appropriate. Difficulties arise as a result of the association of age with various more nebulous statuses such as ‘middle-aged’ and ‘old’. The implications for the concept of ‘the third age’ are discussed and it is concluded that social gerontology should pay more attention to the theoretical significance of chronological age and age-identity and less to age statuses.