The marine microbial community, consisting of autotrophic and heterotrophic bacteria and protists (< 200 µm), is challenging to study due to its high diversity, various trophic functions, and poorly resolved taxonomy. However, the inherent property of organism size, determined by microscopical and flow cytometric techniques, can be used to develop population and community size spectra, which summarize large amounts of information into a simplified format. Illustrating the process with data from the equatorial Pacific, we derive smoothed size density functions for key functional groups based on a relatively new technique from kernel density estimation theory. Optimal bandwidths for each plankton group are determined by a kernel smoothing Bayesian algorithm, and parametric size distribution functions are optimized from the nonparametric kernel density estimates. Resulting density functions are examined for system-level variability in community depth structure and the size relationships of auto- and heterotrophs. When fully developed, we will apply such rigorously defined spectra in the CCE LTER program to evaluate temporal and spatial variability in ecosystem properties, to analyze community responses to experimental manipulations, and to develop and test size-based trophic models.