Nutrient manipulation experiments were conducted on a natural planktonic community in outdoor mesocosms. Inorganic nitrogen (N) and silicon (Si) were added to achieve N[ratio ]Si ratios of 1[ratio ]1 and 4[ratio ]1. Total particulate carbon (PC) biomass of the microbial assemblage was determined by elemental analysis. Cell volume measurements by microscope on individual components of the community (bacteria, diatoms, photosynthetic nanoflagellates, heterotrophic nanoflagellates, dinoflagellates and ciliates) were also made. We applied published C[ratio ]volume relationships to determine the volume estimated C content (CBV) of these microbial groups and hence of the total assemblage. The total CBV and total PC were compared to test the applicability of C[ratio ]volume relationships under different nutrient regimes both before and after nutrient exhaustion. For initial N[ratio ]Si ratios of 1[ratio ]1, prior to nutrient exhaustion, the relationship between CBV and PC was linear with a gradient of approximately 1, (0.99±0.06), indicating that the published C[ratio ]volume relationships accurately predicted the C content of the microbial assemblage. For N[ratio ]Si ratios of 4[ratio ]1, a linear relationship was again evident between CBV and PC (slope: 1.36±0.08). However, statistical comparison using a general linear model indicated that the gradient of this relationship differed significantly from that when the N[ratio ]Si ratio was 1[ratio ]1, and hence CBV overestimated elemental C. For both N[ratio ]Si ratios, subsequent to nutrient exhaustion (N or Si), and hence when the diatom fraction of the microbial assemblage was in yield-limited post-exponential phase, the two measures of biomass were not well correlated. This indicated that measured cytoplasmic cell volume was a poor indicator of C biomass within the microbial assemblage in nutrient-deplete conditions.