The effects of human trampling on a mid-intertidal assemblage dominated by the fucalean alga Hormosira banksii were tested experimentally on two rocky intertidal platforms on the east coast of the South Island of New Zealand. The major factors were Trampling intensity, Platforms and Timing of trampling. Seven trampling intensities (0, 10, 25, 50, 100, 150 and 200 passages per area) were used in 2-m long transects (0.6 m(2)) during single tides in Spring and Autumn on each platform. H. banksii initially had >96% canopy cover in all treatments. As few as 10 tramples reduced this cover by up to 25% after a single tide. Progressively greater reductions occurred at higher trampling intensities, with >90% of the H. banksii canopy removed at 200 tramples. There was recruitment into the Spring transects during the following summer months, and by five months after trampling all treatments had at least 50% cover of H. banksii. Treatments initiated in Autumn had similar reductions in H. banksii cover as Spring treatments, but the recovery was delayed by seven months until recruitment occurred during the following summer. Fifty percent recovery of the canopy took at least a year in most Autumn treatments. After 21 months the Spring treatments had recovered to control levels (>97% cover), while after 16 months the Autumn treatments were still recovering. Understory encrusting and turfing coralline algae were reduced in cover after trampling, with the greatest effects at the higher trampling intensities. The reduction of corallines was mostly due to burn-off after the H. banksii canopy was removed. By the end of the experiment, corallines had returned to control levels in both the Spring and Autumn treatments. Bare primary space increased significantly with up to 60% bare rock at the higher trampling intensities one year after trampling. By the end of the experiment, bare space was reduced to <10% in the Spring treatments but averaged as high as 24% in the higher trampling intensities of the Autumn treatments. There was an interaction between demographic processes, particularly recruitment, and coralline algae in the recovery of H. banksii. Up to 70% of cover at the 200-trample intensity was from new recruits, while most recovery in the 10 and 25 trampling intensities was from re-growth of damaged fronds. However, at one platform, composed of soft siltstone, recruitment of H. banksii occurred almost exclusively on the damaged encrusting and turfing coralline algae. Overall, this experiment showed that trampling intensity had variable effects and that the interaction of season, location, indirect effects of reduction in coralline algae, and facilitative processes in recruitment of H. banksii all contribute to recovery after disturbance. (C) 1999 Elsevier Science B.V. All rights reserved.