No abstract

Many island bird species have been driven to extinction by introduced predators. Although poorly understood, these extinctions could have a 2-fold impact on bird-plant mutualisms, because island bird species can serve as both pollinators and seed dispersers. We investigated how avian translocations into a mammal-free reserve in New Zealand affected the structure of bird-flower and bird-fruit interactions. We observed bird-fruit and bird-flower interactions over a 9-year period to establish (1) the extent to which native birds are both nectivorous and frugivorous (i.e. "dual mutualists") and ( 2) how avian translocations and conservation reestablished nectivory and frugivory networks. Results showed that all but one native bird species were dual mutualists. Pairwise species interaction frequencies were positively correlated between networks. However, overall levels of nectivory by each bird species were unrelated to levels of frugivory. Interaction specialization and species strength also did not differ between networks. The reintroduction of threatened and endangered bird species appeared to have restored both interaction networks, and the sequence of species recovery accelerated restorative changes. Overall results indicate that not only does the extinction of dual mutualists have a 2-fold, negative effect on mutualistic interactions with plants, they can also accelerate the recovery of ecosystem services following restoration efforts.

Many plants produce colour polymorphic fruits. However, the processes responsible for the evolution and maintenance of fruit colour polymorphisms are poorly understood. We investigated the fruit colour polymorphism in Gaultheria depressa var. novae-zealandiae (Ericaceae), a predominantly bird-dispersed, alpine shrub from New Zealand, by testing whether colour morph frequencies vary geographically to maximise fruit-foliage colour contrasts. We also conducted a seed germination experiment to test whether fruit colour morphs vary in their susceptibility to UV damage. Results showed that ‘red’ fruits were more abundant at lower elevations, while ‘white’ fruits were predominant at higher elevations. Leaf colours shifted from ‘green’ in appearance at lower elevations to ‘red’ at higher elevations. Analyses of fruit-foliage colour contrasts showed that ‘red’ fruits were more conspicuous at lower elevations, and ‘white’ fruits being more conspicuous at higher elevations, which was consistent with the hypothesis that colour morph frequencies vary geographically to maximise their conspicuousness to dispersers. However, ‘red’ fruits were generally more conspicuous than ‘white’ fruits, regardless of elevation, indicating that the maintenance of the polymorphism could not be attributed to fruit-foliage colour contrasts alone. The seed germination experiment showed that ‘white’ fruits were more resistant to UV damage, suggesting the preponderance of ‘white’ fruited individuals in the landscape results from a greater degree of protection from UV damage. The fruit colour polymorphism in Gaultheria depressa var. novae-zealandiae therefore appears to be maintained by trade-offs between conspicuousness to dispersers and tolerance to UV damage, advocating a pluralistic approach to the problem in the future.

<p><b>Throughout their evolutionary history, animals have interacted with and evolved alongside plants. There is a long tradition of studying mutualistic relationships between plants and animals. However, the study of mutualistic relationships still present substantial challenges. For instance, the role of dual mutualists in the rebuilding of degraded ecological networks is poorly understood. Additionally, while syndromes have been studied for years, support for the role of fruit and flower colour and their association with animals are still being debated. Finally, even after decades of study, it is still unclear how colours are adaptive in fruit colour polymorphisms and the conditions that lead to them being maintained in nature. It is the goal of this thesis to investigate these understudied aspects of ecology using model species in New Zealand. </b></p> <p>To better understand how dual mutualists influence network properties in New Zealand, I compared how network properties have changed with reintroduction of native bird species. Rate of restoration of network properties was directly impacted by the order of reintroduction. Reintroducing dual mutualists can accelerate the rewiring of networks, but their loss will also have a two-fold impact on networks. </p> <p>Next, I examined how fruit reflectance properties influence frugivore behaviour. Results were consistent with a growing body of research that suggests that fruit reflectance properties are unrelated to frugivore behaviour. On the other hand, research with large scale study systems are uncovering links between fruit colour and frugivore behaviour. Support remains equivocal and may be dependent on scale.</p> <p>Thirdly, I examined whether sexual selection theory can predict flower traits by looking at how a suite of flower traits such as nectar production, flower conspicuousness and flower size are related. Flower traits such as nectar production were correlated with flower size and flower hue. These traits varied between gender consistent with sexual selection theory. </p> <p>Lastly, I investigated how fruit colour in a fruit-colour polymorphic plant can be attributed to both biotic and abiotic conditions. ‘Red’ fruits were more conspicuous than ‘white’ fruits and found at elevations that coincide with higher densities of avian frugivore. On the other hand, ‘white’ fruits were more common at higher elevations, reflected more UV and were better able to protect their seeds from UV damage.</p> <p>Together, these studies broaden our understanding of animal-plant interaction in New Zealand and the importance of dual mutualistic bird species to ecological networks. It provides evidence to a growing body of research on the role of colour in syndromes. Additionally, the results offer new insights into how different conditions influence fruit colour across an elevation gradient.</p>

No abstract