- For many species of oceanic dolphins, photo-identification and genetic data indicate that these island-associated populations are demographically isolated from pelagic populations and that island-associated populations exhibit very different patterns of movement and habitat use. Melon-headed whales (Peponocephala electra) are generally considered a pelagic dolphin, but have been documented around oceanic islands where deep water occurs close to shore. In the main Hawaiian Islands, for example, analysis of ten years of photo-identification data shows that melon-headed whales form a relatively small and apparently isolated population. Melon-headed whales have also been documented around several islands in the Marquesas archipelago of French Polynesia, but no information on their population structure in the islands exists. To better understand this population of melon-headed whale, I analyzed over 6,000 photographs collected during Expedition Marquesas, a series of small-boat sampling surveys conducted around six islands – Hiva Oa, Tahuata, Mohotani, Ua Huka, Ua Pou and Nuku Hiva – in March and April of 2012. Photographs and biopsy samples were collected during 14 encounters with melon-headed whales at four of the islands; Mohotani, Ua Huka, Ua Pou, and Nuku Hiva. In this thesis, I compiled the first photo-identification catalog for melon-headed whales in the islands by analyzing photographs to identify individual dolphins based on distinctive markings to their dorsal fins (referred to as distinctively marked individuals or DMIs). In total, 393 DMIs were identified and sightings were reconciled within encounters, between encounters, and between islands to construct sighting histories for each DMI. This information was then used to improve upon visual counts of the number of individuals in encountered groups, to investigate inter-island movement of DMIs, and to provide a minimum abundance of melon-headed whales in the islands. Estimates of the number of individuals in resting groups were obtained by incorporating information on the number of quality images of each DMI and the total number of quality non-DMI images collected within an encounter into the Poisson lognormal mark-resight model (PNE). The PNE estimates were then compared to visual counts of the number of melon-headed whales in encountered groups obtained by observers in the field. Using the PNE to estimate group size is novel since most studies determine group size by estimating the number of visible individuals. Investigations into the accuracy of observer estimates have found that estimates can vary widely between observers and tend to underestimate group size, especially as group size increases. Here, I found that the sizes of resting groups in the Marquesas Islands were large (median = 458) and that, on average, the mark-resight model estimated the true group size to be 1.5 times greater than the visual counts. By comparing sightings of DMIs between islands, I found evidence of limited movement of DMIs between the surveyed islands. Eight DMIs first seen at Ua Pou were resighted from 3 to 13 days later at Nuku Hiva, a distance of approximately 50 kilometers. No DMIs first documented at Mohotani or Ua Huka were resighted at any of the other islands. While the short time between surveys limits the conclusions that can be made, these results suggest an interesting pattern of habitat use within the islands, especially given the similar distance between encounters at Ua Huka and Nuku Hiva. The minimum abundance of melon-headed whales in the Marquesas Islands was estimated using the Lincoln-Petersen, two-sample, mark-recapture model. I first compared DMI images from pairs of resting-group encounters at each island to identify resightings of DMIs between encounters. Information on the number of DMI images per encounter and the number of DMIs seen in the first encounter and subsequently resighted in the second encounter were incorporated into the model to derive an estimate of the total number of DMIs for each island. Since only information on DMIs was incorporated into the model, the minimum abundance at each island was calculated by adjusting the estimate of the total number of DMIs by the proportion of quality DMI images in each data set. The minimum abundance at each island ranged from 300 (95% CL 188-412; CV = 0.19) at Ua Pou to 631 (95% CL 475-788; CV = 0.13) at Mohotani. Summing the estimates obtained at each island resulted in a minimum abundance estimate of 1,785 (95% CL 1506-2065; CV = 0.36) melon-headed whales in the Marquesas Islands during the study period. The establishment of the first melon-headed whale photo-identification catalog for melon-headed whales in the Marquesas Islands lays a foundation for future investigation into the population structure of this enigmatic species in the islands. The waters around the Marquesas Islands supports a diverse marine community, including several species of small cetaceans, and represents an area of high productivity in an otherwise oligotrophic region of the South Pacific. French Polynesia has taken steps to protect this important habitat by designating the entirety of the French Polynesia EEZ as a Marine Mammal Sanctuary and by petitioning UNESCO to designate the Marquesas Islands as a World Heritage Site. The baseline information provided by this thesis is an important step in understanding the population structure and habitat use of melon-headed whales – information crucial to effective conservation of this species in the Marquesas Islands.