COS 105-3
An integrated framework to improve the concept of resource specialization
Resource specialization is a fundamental element of ecological theory. However, it is defined and evaluated in very disparate ways. In most cases, operational definitions are variants of a simple count of resource classes, most often of species. Here we build on recent advances to propose a concept of specialization that comprises affinities among resources as well as their co-occurrence with consumers. In our null modeling approach, specialization is measured as phylogenetic relatedness of trophic associates, scaled by the null expectation of random use of locally available partner species. Hence, specialists are species that use significantly clustered phylogenetic sets of resources, whereas hiper-generalists feed on significantly over-dispersed resources. Intermediate species are then classed as indiscriminate feeders. To ascertain the effectiveness of this approach, we constructed increasingly restricted null models. We then applied it to a dataset of 168 species of herbivores on Brazilian Asteraceae flowerheads belonging to four families.
Results/Conclusions
Within our framework, we were able to classify the herbivore species as specialists, generalists or indiscriminate feeders, with marked differences among the herbivore families regarding the proportion of species in these specialization categories. Incorporation of plant relatedness and relative abundance greatly improved specialization measures compared to taxon counts or simple null models, which tend to overestimate the fraction of specialists. With this approach the concept of specialization is disambiguated and its measure becomes explicit and rigorous, directly including ecological and evolutionary aspects. We expect that this framework will enable a more rigorous usage of the concept of ecological specialization in empirical and theoretical studies.