Degree Name

Doctor of Philosophy


School of Biological Sciences


Group living behavior has evolved in every major taxonomic classification including plants, bacteria, birds, mammals, insects, reptiles and fish. The widespread nature of this behavior is inherently interesting because the mechanisms behind its evolution are highly variable, even between closely related species, and are not always immediately obvious. Cooperative breeding theory was developed to study groups in which subordinates cooperate to help raise the offspring of dominant breeders. The hypotheses which make up cooperative breeding theory are also useful for examining the evolution of sociality in groups with social systems other than cooperative breeding. Thus cooperative breeding theory provides a rich framework with which to assess social evolution in an extensive range of species. Cooperative breeding theory was initially developed based on data from terrestrial species living in family groups (that is, with high kinship). In such groups, helpers may gain indirect benefits from helping to raise the offspring of related breeders. However, a number of taxa are now known to form groups of unrelated individuals and the genetic benefits conferred to helpers are either substantially reduced or non-existent. There must, therefore be alternative benefits which outweigh the costs (e.g. delayed breeding) of group living. The initial focus on terrestrial taxa has lead to a taxonomic bias in the literature, tilted toward terrestrial animals. While great advances have been made in our understanding of the factors involved in social evolution using these model terrestrial systems, we are, nonetheless limited to understanding sociality through a terrestrial lens. Aquatic taxa have received far less attention in the sociality literature as they have historically been overlooked as either non-social or only exhibiting very basic group-forming behavior. A good body of research has since been conducted on freshwater cichlids and this has resulted in a much broader comprehension of social group evolution and maintenance. However, cichlids still conform to the terrestrial paradigm of congregating in family groups. Many marine fishes on the other hand, have a pelagic larval phase. This life-history characteristic means that groups of these fishes generally have low relatedness. Other life-history characteristics such as sex-change are commonplace in many lineages of marine fishes and are rarely observed in terrestrial taxa. Such abilities conceivably alter the various costs and benefits of group living, making marine fishes a very interesting model system for social evolution and maintenance. In chapter 2, I therefore aimed to reduce the taxonomic bias in the literature by investigating the evolution and maintenance of sociality in habitat-specialist coral reef fishes (genus Gobiodon). These fishes are demersal spawners with a pelagic larval phase and thus are suspected to form non-family groups. I achieved this aim by first demonstrating the taxonomic bias present in the field through a critical review of the literature on social evolution. I also suggest a cohesive framework with which to progress research in this field. I then use this framework in the following chapters to address the taxonomic bias by examining social evolution and providing a solid foundation for future research in a new model system of marine fishes.

In chapter 3, I set the foundation for future work on social evolution in the genus by providing a multifaceted description of sociality in each species of Gobiodon present at Lizard Island, Australia. I collected data on group sizes, host-coral (habitat) sizes and fish length and took tissue samples for genetic analyses from each species of Gobiodon. The vast majority of studies examining sociality use group size as a metric for sociality. My description of sociality involved quantifying sociality by two methods (group size and a more complex sociality index), examining social structure within groups, assessing constraints on group size and investigating the distribution and ancestral states of sociality throughout the genus. I verified group size as a reasonable proxy for sociality in Gobiodon, which, to my knowledge, has not been attempted in any other study of sociality. I also found there was good evidence for size based hierarchies in the more social species of Gobiodon by regressing fish length against size rank. In the analysis of constraints on group size, coral size (habitat) was a significant predictor of group size in the majority of the group-forming goby species. Lastly, I built a phylogeny specific to the Gobiodon species at Lizard Island and reconstructed the ancestral states of sociality. I found that sociality was randomly distributed throughout the genus suggesting that factors other than phylogenetic constraint were likely responsible for the evolution of sociality in the genus. I reinforced this finding in chapter 4, by conducting an analysis of phylogenetic signal of sociality in Gobiodon, which provided ambiguous results. These results did however suggest that if there was a phylogenetic signal of sociality in the genus, it was weak and likely had little effect on social evolution. I therefore conducted an examination of the evolution of sociality in the genus with phylogenetically controlled comparative analyses of ecological (coral size and host generalization) and life-history (average body size) traits. I found a combination of coral size and body size best predicted sociality in the genus.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.