Plumage polymorphism is displayed in a number of diverse species and therefore, is presumed to have evolved independently on multiple occasions. To date, no study has established whether any of the hypotheses proposed to explain its evolution in single species are relevant across all avian taxa. Using comparative analyses which included all species displaying a plumage polymorphism involving melanin pigmentation, I found no support for either sexual or apostatic selection being responsible for the evolution of plumage polymorphism. Weak support for a role for disruptive selection was indicated in certain species but not when all polymorphic species were considered together. Contrary to previous studies, this indicates that plumage polymorphism may not confer any selective advantage but may in fact be selectively neutral or perhaps simply a product of historical biogeographic processes. Many polymorphic species have clinal morph-ratio distributions. In the polymorphic New Zealand fantail, Rhipidura fuliginosa, I established that the black morph was more common in central parts of the South Island of New Zealand than in the South. This was not the pattern expected based on previous studies and no environmental factor was found to correlate with the distribution. Polymorphism may not confer a selective advantage to a species as a whole, but the fact that two or more often very differently coloured morphs can persist within a population suggests that balancing selection may be in operation. I demonstrated that black morph fantails suffered reduced feather damage over the course of the year and that black and pied fantails differed slightly in foraging behaviour. The benefits of black and white plumage whilst foraging were further elucidated through plumage manipulation experiments. Thus, a trade-off between feather wear and foraging between the two morphs of the fantail, produced due to the different selective advantages provided by their plumage colouration, seems likely to be part of the balancing selection mechanism responsible for maintaining the plumage polymorphism in this species. Plumage polymorphism may be better understood if the mechanisms controlling plumage evolution in general were unravelled. Therefore, I reconstructed plumage characters of the genus Rhipidura onto a molecular phylogeny that I created based on the cytochrome b gene. Species that were divergent in plumage were also found to be more genetically divergent. This suggested that, unlike in other genera, the plumage characters of the Rhipidura spp. may be a good estimator of phylogenentic relationships.