Abstract Several cage experiments were conducted in the shallow reef along the south-west coast of Curaçao, in order to test the hypothesis that “undergrazing” might cause degradation of the coral reef as a consequence of unhampered growth of fleshy and filamentous algae covering the important reef building and consolidating crustose corallines and thus causing their death. Substrates excluded from grazing were covered rapidly by a conspicuous vegetation of filamentous and fleshy algae. In this respect, differences were found neither between artificial substrates (asbestos plates) and natural substrates (crustose coralline algae and dead coral limestone) nor between substrates whose surfaces were initially devoid of filamentous and fleshy algae (asbestos plates and crustose coralline algae) and substrates which bore a sparse vegetation of epilithic algae (dead coral limestone). The reef building and consolidating crustose corallines ( Porolithon pachydermum (Fosl.) Fosl. and Lithophyllum sp.3) died when excluded from grazing, probably because they were covered by a dense vegetation of filamentous and fleshy algae. The superficial layer of the dead coralline crusts were infested by endolithic algae. Crustose coralline algae were only able to settle and grow on substrates which were not excluded from grazing. Grazing by fishes alone was sufficient to suppress increment of epilithic, fleshy and filamentous algae. Gross production of the dead coralline crusts, with their endolithic and epilithic algae (0.028–0.046 mg O 2/cm 2/h) was as high as gross production of the living coralline algae. It is concluded that the high grazing pressure exhibited by the numerous herbivores, does not affect the high primary productivity of the shallow reef system seriously. When a natural sparse vegetation of filamentous and fleshy algae was excluded from grazing, it developed into a dense vegetation with a species comoosition and gross production which was comparable to the natural dense vegetation inhabiting crevices and hollows in the shallow reef. However, the natural dense vegetation, which is grazed to some extent has a higher productivity/biomass ratio than the dense vegetation completely excluded from grazing. The importance of endolithic, boring algae, which are present everywhere in the limestone of the shallow reef is stressed. Whenever an overlying primary producer disappears, its task as primary producer is replaced by the boring algae. The composition of the vegetation which appeared on the artificial substrates showed a distinct succession. On substrates excluded from grazing, during the first 6–8 weeks ephemerous species like Giffordia duchassaingiana (Grün.) Taylor, Enteromorpha clathrata (Roth.) Grev. and Cladophora spp. were important constituents, and after a prolonged period (10–15 weeks) more fleshy and larger species (e.g. Centroceras clavulatum (C.Ag.) Mont., Wrangelia argus Mont., Jania capillacea Harv., Dictyota dichotoma (Huds.) Lamour., Laurencia microcladia Kütz., Spyridia filamentosa (Wulf.) Harv., Pterocladia americana Taylor) became abundant. Much sediment was entrapped by these dense vegetations. In the sparse vegetation on the plates where fishes were allowed to graze, Giffordia duchassaingiana, Enteromorpha clathrata and Trichosolen sp. were abundant at first. However, on these plates, probably because they were grazed, several crustose algae ( Porolithon pachydermum, Neogoniolithon solubile (Fosl. et Howe) Setch. et Mason, Protoderma sp. and Myrionema sp.) were able to settle and become important constituents of the vegetation.