1. 151 woods, ranging in size from 0.02 to 30 ha and with differing de
grees of isolation, were surveyed to determine the numbers of breeding
bird species present in each of three consecutive years. Woodland and
edge species were analysed separately. 2. Numbers of woodland species
showed a strong relationship with woodland area which explained c. 70
% of the variation in breeding species numbers. Several other features
of the woods and of the surrounding landscape also had small addition
al influences on species numbers. 3. Numbers of edge species showed a
poor relationship with woodland area, but length of perimeter explaine
d 30-40% of the variation in species numbers. The number of structural
features within a wood and isolation from other woodland also had sma
ll additional positive influences on edge species numbers. 4. Small wo
ods held more edge species than equivalent areas of large woods, but t
ended to hold fewer woodland species than equivalent areas of large wo
ods.