Microbial size spectra, including bacteria through nanophytoplankton, were measured by use of flow cytometry across the western north Atlantic Ocean and during two nutrient enrichment studies: bottle enrichments in the Sargasso Sea and an in situ iron enrichment in the equatorial Pacific (IronEx II). Spectral shapes, or the relative conformity to a function described by a power law, ranged from smooth and log linear during the spring bloom in the Sargasso Sea to being distinctly non-log linear in coastal waters. Overall, the individual spectra within large regions characterized by similar ecological conditions showed remarkable consistency, inviting speculation that powerful organizing mechanisms are at work in these communities. Moreover, the ensemble average of all of the spectra along the transect displays clear power-law behavior. Slopes ranged from Ϫ1.0, in which biomass was equally distributed between all size classes, to Ϫ1.4, in which proportionally more biomass was contained in smaller size classes; there was no clear relationship between nutrient concentrations and spectral slopes over the entire data set. Species succession in nutrient-enriched bottles caused spectra to evolve from relatively smooth power laws to distributions showing preferred sizes (i.e., nonlinear on a log-log plot). The IronEx II spectra, however, remained similar over the course of the experiment. It could be that the elimination of bottle effects in this experiment buffered the system in ways that maintained the size structure of the microbial community over the size range we measured. Our results suggest conditions that lead to log-linear size distributions; these should be verified over a broader range of scales and environments.Size spectra, which display the relative abundance of organisms of different sizes, convey a synoptic image of ecological communities that is taxon independent. As such, they have been attractive to ecological theorists and have been the subject of periodic interest in marine ecology for the past 30 yr. Sheldon et al. (1972Sheldon et al. ( , 1977 recognized the predictive powers of size spectra, suggesting that fish stocks could be predicted if the planktonic size spectrum were known. Moreover, a spectral approach offers potential for enhancing ecosystem models (Gin et al. 1998)
AcknowledgementsWe thank the captain and crew of the R/V Oceanus, who facilitated the collection of the transect data through challenging weather conditions. We thank M. Durand and R. Greene for help developing the Mie fit to our calibration data. We also thank R. Olson and L. Moore for helpful comments on earlier drafts of this manuscript and J. Rodríguez and an anonymous reviewer for valuable critical comments.