Assuming heterotrophic denitrification as the dominant microbial process, Richards (1965) formulated a stoichiometry governing nitrogen loss in open-ocean oxygen deficient zones (ODZs). It prescribes the quantitative coupling between the oxidation of organic matter by NO3 in the absence of O2 and the corresponding production of CO2, N2, and PO–34. Applied globally, this relationship defines key linkages between the C, N, and P cycles. However, the validity of Richards's stoichiometry is challenged by recognition of complex microbial N processing in ODZs including anammox as an important pathway and nitrite reoxidation. Whereas Richards's stoichiometry would result in N2-N production to NO3 removal rates of 1.17, dominance by anammox with respect to biogenic N2 production could in theory result in a ratio as high as 2. Ratios with PO–34 production provide an additional constraint on the quantity and composition of respired organic matter. Here we use a mesoscale eddy with extreme N-loss in the Peru ODZ as a "natural laboratory" to examine N-loss stoichiometry. Its intense biogeochemical signatures, relatively well-defined timescales, and simplified hydrography allowed for the development of strong co-occurring gradients in NO3, NO2, biogenic N2, and PO–34. The production of biogenic N2 as compared with the removal of NO3 (analyzed either directly or as N deficits) was slightly less than predicted by Richards's stoichiometry and did not at all support any "excess" biogenic N2. PO–34 production, however, was twice the expectation from Richards's stoichiometry suggesting that respired organic matter was P-rich as compared with C:N:P Redfield composition. These results suggest major gaps remain between current understanding of microbial N pathways in ODZs and their net biogeochemical output.