This thesis examines methodological issues in the analysis of fishbone assemblages from the Pacific. A review of the literature on Pacific fishing raises questions about methodology in Pacific fishbone analysis. Recent work in Pacific fishbone analysis at the Otago Archaeology Laboratory and other international laboratories has shown that different identification and quantification methods have significant influence on measures of relative abundance. This can affect the type of interpretations that result from midden analysis. A common method employed in New Zealand and the Pacific involves the identification of five paired mouth bones plus various special bones, and the use of MNI for quantification. A large fishbone assemblage from Rapa in French Polynesia is used as a case study to evaluate alternative analytical approaches. Tropical Pacific fish bone assemblages tend to have a relatively high species diversity compared to temperate New Zealand assemblages. Many reef fish also have small mouths in comparison to temperate fish species. It is concluded that in tropical Pacific assemblages the number of elements identified to taxa can have a significant effect on possible interpretations of past behaviours.This thesis also presents the first archaeological interpretation of the prehistoric fishing on Rapa. Due to its subtropical location Rapa presents an opportunity to study human adaptation to resource poor environments. While it is found that the fishing industry shows some general similarities to those of other islands in East Polynesia, the environmental constraints have resulted in some interesting variations in terms of target taxa.The analysis of the Rapan assemblages confirms that, at least in some cases, the use of a wider suite of elements for identification will have a significant impact on the number of taxa identified in an assemblage, and resulting interpretations of fishing behaviours.