An ecosystem engineer's influence on associated organisms includes 2 components: physical and biological. The physical component is the spatial environment created by the ecosystem engineer, while the biological component is the physiological activity of the engineering species. However, little is known about the ratio between these components. We evaluated this ratio for Mytilus edulis L. by means of a field experiment in which communities of organisms that developed on bare ceramic plates (controls, C) were compared with those that formed on plates carrying patches of live mussels (L) or artificial patches made of mussel dummies (D). The experiment was performed using 2 different age-size groups of mussels and dummies (small-young and large-old). Live mussels had only a weak positive effect on species richness of the associated organisms, but the structure of the communities that formed on tests plates (L, D and C) differed significantly. The similarity of the communities of associated organisms between C and L was 1.5-3 times less than between C and D. Mussel size-age factor had no significant effect on community structure other than on the algal assemblage. The results of this experiment suggest that the influence of mussels on associated organisms cannot be reduced to only the effect of increasing complexity of the spatial environment. The influence of the biological component is significant and can exceed that of the physical component. Thus, in ecosystem engineering, non-living spatial structures cannot serve as an identical substitute for a population of living organisms.