Fisheries on coral reefs are highly complex, can be very productive, b
ut typically have little or no management. Widespread overfishing and
declining yields reveal an acute need for proper management. However,
conventional management methods are inappropriate for two main reasons
: they require much information on the biology of stocks and are expen
sive and difficult to enforce. Use of marine reserves has been suggest
ed as an alternative. Protective management potentially has several im
portant benefits including (i) protection of spawning stocks; (ii) pro
vision of recruits to replenish fishing grounds; (iii) enhancement of
catches in adjacent unprotected areas through emigration; (iv) minimal
requirement for information on biology of stocks; and (v) ease of enf
orcement. However, the effectiveness of the reserve approach has not b
een properly tested. We evaluate the evidence available to test whethe
r reserves function as predicted on theoretical grounds. In general, f
ield studies from widespread sites around the globe support prediction
s of increases in abundance and average size of fishes in protected ar
eas. However, evidence for enhanced catches in adjacent areas is more
limited, and evidence to show that reserves can restock fishing ground
s is lacking. Nevertheless, protective management appears to hold much
promise for low-cost management of reef fisheries. Research programs
in several areas of the Caribbean and Indo-Pacific have now been launc
hed to refine the approach.