Arthrospira (Spirulina) platensis (Nordstedt) Gomont was cultivated under light-limited conditions in 5-L open tanks by daily supplying NH 4 Cl as nitrogen source. Exponentially increasing feeding rates were adopted to prevent ammonia toxicity. The total feeding time (T) was varied between 12 and 20 days, and the starting (m 0 ) and total (m T ) quantities of the nitrogen source per unit reactor volume were varied in the ranges 0.19-1.7 mM and 2.3-23.1 mM, respectively. This intermittent addition of the nitrogen source prevented ammonia from reaching inhibitory levels and ensured final cell concentrations (X m ) and cell productivities (P x ) comparable with those of batch runs with KNO 3 . Moreover, the lower nitrogen addition due to the use of NH 4 Cl rather than KNO 3 allowed for higher nitrogen-to-cell conversions (Y x/n ). These results were evaluated using three-factor, five-level, central composite experimental planning, combined with the response surface methodology, selecting T, m 0 , and m T as the independent variables and X m , P x , and Y x/n as the response variables. This approach allowed us to identify, through the simultaneous optimization of the variables, T 5 16 days, m 0 5 1.7 mM, and m T 5 21.5 mM as the best conditions for A. platensis cultivation at 72 lmol photons . m À 2 . s À 1 . Under these conditions, a maximum cell concentration of 1239 mg . L À 1 was obtained, which is a value comparable with that obtained using KNO 3 as nitrogen source and nearly coincident with the theoretical one estimated by the response surface methodology.