I write in reply to the letter by Combridge et al. (2003), and to agree with their view that the claim by Roberts et al. (1999), of `a real increase' in numbers of breeding Honey-buzzards Pernis apivorus in Britain in the last 30 years, has not yet been substantiated. Since 1954, a small group of observers has monitored Honey-buzzards in the New Forest, Hampshire, and their records (summarised for 1954-92 in Clark & Eyre 1993) show that the number of pairs present in the breeding season has fluctuated. Some allowance for inexperience in monitoring the species during the early years of fieldwork in the 1950s needs to be made, but up to four pairs were present during 1954-60, up to nine pairs during 1961-80 and two or three pairs during 1981-92. The period 19932002 saw a recovery to pre-1981 population levels, although in 2003 numbers roughly halved and were then similar to those during 1981-92. It is worth stressing that the above figures include pairs which summered but did not attempt to breed. In my view, it is perfectly reasonable to suggest that the British population as a whole may have been subject to similar fluctuations, though with so many variables affecting Honey-buzzards (e.g. numbers reaching Britain, regional climatic differences in any given year), such variations may not be consistent throughout the country. Honey-buzzards are long-lived birds (one is known to have reached 28 years 10 months; BWP) and individual pairs can be remarkably successful, occupying a site for a