Macroevolutionary Changes of Plants on Islands

<p>Insularity is known to produce predictable evolutionary changes in plants. For example, herbaceous plants often evolve woodiness and seeds tend to have reduced dispersal capabilities on islands. However, our understanding of how other plant traits may evolve on islands is lacking. Furthermore, plants are modular organisms and by investigating evolutionary changes in specific plant traits we may better understand macroevolutionary processes on islands.  In this thesis, I investigate evolutionary changes in a range of plant traits on islands. First, I tested for evolutionary changes in seed size on islands (Chapter 2). Island plants consistently produced larger seeds than mainland relatives. Furthermore, this result was consistent regardless of differences in dispersal mode, growth form and evolutionary history. Selection may favour increased seed size to reduce dispersal distances. Additionally, selection may favour larger seeds due to the competitive advantage conferred to developing seedlings.  Many animal taxa exhibit increased sexual size dimorphism (SSD) on islands, as predicted by the niche variation hypothesis. However, patterns of SSD among dioecious plants on islands are unknown. In Chapter 3 I tested for differences in SSD of dioecious plants that colonized four island groups from New Zealand (mainland). The degree of SSD did not vary predictable between island and mainland plants, contrary to predictions of the niche variation hypothesis. However, SSD was consistently female biased on the mainland and results suggest selection is acting to increase the size of both sexes on islands.  Evolutionary changes in island plants may be a response to herbivory by unique large browsers. For example, the divaricate growth form is common in the New Zealand flora and may have deterred browsing moa. In Chapter 4 I tested for differences in traits associated with the divaricate growth form between plants from mainland New Zealand and Chatham Island. Results suggest that an absence of moa on Chatham Island has relaxed selection for traits associated with the divaricate growth form.  An emerging body of research suggests aposematism (warning signals to herbivores) may be common in plants. However, previous investigations have not appreciated the fact that the perspective of terrestrial herbivores changes as plants grown vertically. Furthermore, ontogenetic changes in the capacity of plants to defend themselves may influence the reliability of warning signals. In Chapter 5 I tested for ontogenetic changes in two potentially aposematic signals produced by Pseudopanax crassifolius. Aposematism on upper leaf surfaces peaked early in ontogeny, providing a dishonest signal of defense. Conversely, signaling on the underside of leaves peaked later in ontogeny and scaled positively with structural defenses.   The results of this thesis suggest selection is acting on specific plant traits on islands. Evolutionary pathways, such as the evolution of woodiness, may be better explained by considering selection acting on other plant traits. For example, selection acting on seed size may facilitate evolutionary size changes evident at later life-history stages. A lack of consensus exists regarding the role of insular herbivores in the evolution of island plants. The results of Chapters 4 and 5 suggest herbivory has played an important role in the evolution of novel morphology of island plants. Considering trait specific changes of plants on islands may further our understanding of prominent evolutionary pathways by pinpointing the action of selection.</p>