Recent letters to British Birds question whether there has, in fact, been a real increase in breeding Honey-buzzards Pernis apivorus in Britain (Wiseman 2004), and suggest that the present population is `not certainly more than 30-50 pairs' (Mummery et al. 2004). Over the past 20 years, I have watched breeding Honeybuzzards at more than 30 locations in England, Scotland and Wales, and have discussed their present and past status with fieldworkers in the relevant areas. In any discussion of population numbers, for any species, it is important to establish just what `accounting unit' is being used. Obviously, there are likely to be more territorial pairs in a population than there are successful breeding pairs, especially for a large raptor such as the Honey-buzzard. Contrary to the quoted experience of Mummery et al. in the New Forest, Hampshire, a significant proportion of birds present at most British Honey-buzzard sites are non-breeders. Several long-term studies in The Netherlands have found that less than 50% of resident Honey-buzzards were engaged in a breeding attempt in any one year (van Manen 2000; Voskamp 2000; Bijlsma 2004). This high proportion of non-breeders is mirrored in studies of Honey-buzzards in several other European countries (R. Bijlsma pers. comm.). In fact, many of the observations in the letter from Mummery et al. are either at odds with experience in other parts of Britain (e.g. Irons 1980, Roberts et al. 1999), and much published data from other parts of Europe, or are simply not provable. Although the British