The size of an organism is one of its most immediately obvious characteristics, and many key physiological, ecological, and evolutionary processes scale with size. Size is also relevant in the context of size-selective harvest which is common for many organisms. Using size-based approaches I report new insights into the photophysiology and demography of giant clams, to inform management of size selective harvest. First, I found that the photosynthetic performance (photosynthetic efficiency and relative electron transport rates) of symbiotic Symbiodinium in mixotrophic giant clams (Tridacna maxima) varied as a quadratic function of clam size. Second, I modeled a size-selective, artisanal fishery for T. maxima at a spatial scale (the island of Moorea, French Polynesia) where the amount of self-recruitment is unknown. I collected demographic data from 12 sites on Moorea annually from 2006-2010 to construct a size-based integral projection model (IPM) with a mix of self-recruitment and external recruitment. The IPM results indicated that in the worst case scenario, the population of giant clams on Moorea will exhibit a 7% decline in abundance under the present-day fishing mortality rate if it has 0% self-recruitment, but will increase exponentially if it has 100% self-recruitment. Fishing mortality decreased the projected population abundance. Finally, I evaluated the use of minimum size limits to optimize annual harvest across all possible amounts of self recruitment in a local population of giant clams. I used the IPM to project harvest for combinations of minimum size limit and self-recruitment ranging from 0-100% of total recruitment. A single near-optimal size limit of 135 mm resulted in maximum or near-maximum harvest across the range of self-recruitment. A single near optimum size limit can be applied for organisms with a variety of different life history characteristics unless a local population exhibits exponential decline. Overall, I found that the local population of giant clams on Moorea likely can support the present-day fishing mortality rate, and that size limits can optimize or nearly optimize harvest of local populations. My work indicates that populations can be modeled and managed at the local population scale in the face of uncertainty regarding the amount of self-recruitment.