All organisms are faced with a chemical imbalance between their internal environment (cells, tissues, or body) and their external environment. Homeostasis is the ability to maintain an internal state that is different from the external environment and at least some degree of elemental homeostasis is required for metabolism and growth. Homeostasis is related to fitness since the degree of elemental imbalance between an organism's biomass and its resources controls the growth of populations, predicts the outcome of competition, and determines the relative rates of resource consumption, assimilation, and excretion of elements and energy. Since all organisms are composed of molecules that are comprised mainly of a common set of elements (carbon (C), hydrogen, oxygen, nitrogen (N), phosphorus (P), etc.), stoichiometric ratios of these elements in biomass (e.g. C:Pbiomass) and resources (C:Presources) can be used to diagnose the strength of imbalance and to assess the nutritional state of organisms. The strength of elemental homeostasis is variable within and among groups of taxa; some species and groups maintain strong homeostasis, but others adjust their chemical composition in response to their environment. Since ecosystems seldom contain only a single species, assemblages and communities can respond to elemental imbalance both through changes in the relative abundance of species and through simultaneous changes in the elemental content of the component species. The goals of this dissertation are to evaluate the role of resource competition and species shifts in the stoichiometry of assemblages and to understand the ranges of stoichiometric regulation and biomass chemistry within the bacterial assemblages of lakes. In chapter 1, I introduce the conceptual framework of `stoichiometric strategies' to align the gradient of stoichiometric regulation with physiological tradeoffs. Data from previously published studies on planktonic organisms show that the strength of homeostasis in a species is inversely proportional to the ratio of the two elements in its biomass when the denominator element is limiting. Under nutrient limitation, homeostatic species have lower biomass C:N, C:P, and N:P ratios than do species with flexible biomass stoichiometry. I show how a consumer-resource model with tradeoffs related to competitive ability for C and P couples homeostatic regulation to competitive ability. The result is a conceptual model in which assemblages are dominated by homeostatic species under low resource imbalance and by species with flexible stoichiometry when nutrients are strongly limiting. I test the stoichiometric strategies concept in chapter 2 by culturing assemblages of heterotrophic bacteria at a range of resource ratios and examining the strength of homeostasis in the dominant species. I found that low resource C:P ratios could select for homeostatic strains of bacteria and that higher resource C:P ratios yielded assemblages with flexible composition. In chapter 3, I use bacteria isolated from lakes to describe how homeostatic strains and flexible strains respond to imbalance in C and P. The strains exhibited substantial variation in stoichiometric regulation, but strong homeostasis was associated with higher C and P content and flexible stoichiometry was present only in strains with low P content. These experiments support the hypothesis that flexible biomass composition is competitively superior under P limitation. In the final chapter, I seek to characterize the range of cellular P content attainable by heterotrophic bacteria and determine how bacteria minimize their P content in response to P limitation. I show that bacteria can exhibit greater flexibility in P content than was known previously (< 0.01 to 3% of dry mass as P, biomass C:P of 30:1 to > 10,000:1) and that this flexibility is explained by a simultaneous increase in C content (13 to > 70 fmoles cell-1) and decrease in P content (0.62 to < 0.06 fmoles cell-1) under P limitation. </DISS_para> <DISS_para>These studies highlight the importance of physiological constraints and assemblage-level interactions to understanding the impact of stoichiometry on biogeochemical cycles. Additionally, the results of these experiments show that strains of bacteria differ dramatically in their elemental composition, stoichiometric regulation, and resource demands and that the assumptions of strong homeostasis and high nutrient content are not representative of bacteria in aquatic environments. Although aquatic heterotrophic bacteria serve as a useful system to address these questions, the constraints appear to be fundamental and these results are likely applicable to other groups of organisms.