Phytoplankton fluorescence has been used historically as a means of assessing phytoplankton biomass, rates of primary production (PP) and physiological status in laboratory, in situ, and satellite based investigations. Assumptions about the quantum yield of phytoplankton fluorescence, φ f , are often overlooked and can become problematic when fluorescence based methods are applied. A time series of φ f observations from the northwestern Sargasso Sea is presented with the goal of understanding the controls on fluorescence and its applicability for assessing upper ocean biological processes. Accurate estimates of φ f require accounting for Raman scattering and the conversion of planar to scalar irradiance. Variability in φ f occurs on both seasonal and episodic time scales. Seasonal variations show maxima in the surface layer during summer months while lower, more uniform values are found throughout the winter when deep mixing occurs. Large episodic variations in φ f are observed throughout the record which dwarf seasonal changes. Predictions of depth-dependent and depth-integrated PP rates using φ f and natural fluorescence fluxes are only marginally successful (r 2 ~50%), although comparable with results from global bio-optical models for the Sargasso Sea. Improvements in PP predictions are hindered by weak statistical relationships with other parameters. φ f is largely decoupled from the quantum yield of carbon assimilation, φ c , indicating that an inverse relationship between fluorescence and photosynthesis does not exist. Consequently, variability in the quantum yield of thermal de-excitation, φ h , is found to be of similar magnitude as φ f on the timescales observed. These observations show that assumptions about photochemical energy flow through the phytoplankton community must be made carefully and that the fluorescence-photosynthesis relationship is not straightforward.