however, once Fucus became established, percent coverage increased 

 rapidly. By September 1983, Fucus occupied 40-50% of the low intertidal 

 at the sheltered stations. Recovery of Fucus in the raid intertidal of 

 the recolonization stations, and comparisons of Fucus coverage in 

 recolonization versus undisturbed transects since 1979, are shown in 

 Figure 4. 



Another way of illustrating recovery of the intertidal communities 

 after denudation is to calculate similarity coefficients for 

 recolonization and undisturbed transects, and apply the clustering 

 algorithm described above. An example of this technique is shown in 

 Figure 5, where data from the mid intertidal of the autumn denudings at 

 Fox Island-Exposed and Fox Island-Sheltered are presented. At Fox 

 Island-Exposed (Fig. 5a), four clusters are apparent; the first three 

 correspond well with the seasonal groupings described for the entire 

 area, i.e. summer /autumn, winter, and spring (cf. Fig. 3). Cluster D 

 consists of the first three collections made subsequent to denudation; 

 they show little similarity to each other (because of increasing amounts 

 of ephemeral algae) , and even less similarity to any other combination 

 of collections (due to absence of barnacles and Fucus ) . Succeeding 

 collections at the recolonization strips (Apr., May, Jul. 1982) reflect 

 recovery of barnacle populations, but due to scarcity of Fucus, show 

 more similarity to winter collections at undisturbed transects. By 

 September 1982 (12 months after denuding), the recolonization strips 

 closely resemble other autumn collections, and all subsequent sample 

 months cluster 'correctly'. In July and September 1983, recolonization 

 and undisturbed transects were highly similar, and recolonization at 

 this site was considered complete. 



At Fox Island-Sheltered, however, recovery towards pre-experimental 

 conditions has been much slower (Fig. 5b) . At this station, ephemeral 

 algae are much less common than at the exposed stations, so the 

 seasonality of the flora is not as evident. Rather, Fucus is the 

 dominant algal component of the mid intertidal community, and Cluster A 

 represents collections in 1981 and early 1982 when fucoid cover is low. 

 As discussed above, Fucus abundance undergoes natural fluctuations, on a 

 2-4 year cycle, and Cluster B represents the upswing of the fucoid 

 canopy (Jul. 1982-Sep. 1983) in the undisturbed transects. Cluster C is 



19 



