The role of recruitment and behaviour in the formation of mussel-dominated assemblages: an ontogenetic and taxonomic perspective

Behaviour influences individual fitness with effects that can propagate from the individual to the group. Here, we tested for higher-level effects of individual behaviour in the structuring of intertidal populations of two competing ecosystem engineering species. We used the partial habitat segregation exhibited by co-occurring indigenous ( Perna perna ) and invasive ( Mytilus galloprovincialis ) mussels in South Africa to test for possible attraction of different size classes of recruits to conspecific adults, using a combination of field and laboratory studies. Each of the two species dominates a particular height on the shore with overlap in the mid-mussel zone, but measurements of settlement and recruitment in the field partially refuted previous findings, generally showing no within-shore pattern of zonation of settlers and recruits. At smaller scales, recruits of both species were found more frequently on adults of Mytilus in natural beds where adults coexist in mixed-species populations. Finally, the results of laboratory choice experiments showed that recruits of all sizes responded to adult cues by movement, but that the smallest recruits showed only minimal movement and never reached adults; only large recruits of Perna responded positively to conspecific Perna adults. This study emphasises how observations made at different scales, from shore (among sites) to mussel bed (within shores), to the individual (field and laboratory), can produce different, or even contrasting, information, highlighting how behavioural traits, like attraction to conspecifics, can differ within the same group of organisms (congeneric species) and change ontogenetically within a species. Incorporating fine-scale responses makes predictions of population dynamics more complex, but identifying the relative strengths of mechanisms that lead to patterns of distribution is necessary for understanding higher-level interactions within a system.