Environmental context Phytoplankton form the base of marine food-webs, and hence they have been proposed as the likely source of many arsenic compounds found in marine animals. Because of the difficulties associated with field experiments with phytoplankton, attempts to test this hypothesis have relied mainly on laboratory experiments. This study assesses the environmental validity of this research approach by investigating the influence of the culturing experimental protocol on the uptake, accumulation and biotransformation of arsenic by marine phytoplankton. Abstract Arsenic cycling by the marine phytoplankton Dunaliella tertiolecta and the marine diatom Thalassiosira pseudonana was influenced by culture regime. Arsenic was associated with the residue cell fractions of batch cultured phytoplankton (D. tertiolecta and T. pseudonana), due to the accumulation of dead cells within batch cultures. Greater arsenic concentrations were associated with water-soluble and lipid-soluble cell fractions of continuously cultured phytoplankton. Arsenoribosides (as glycerol (Gly-), phosphate (PO4-) and sulfate (OSO3-)) were ubiquitous in D. tertiolecta (Gly- and PO4- only) and T. pseudonana (all three species). Additionally, arsenobetaine (AB) was not detected in any phytoplankton tissues, illustrating that marine phytoplankton themselves are not an alternate source of AB. Arsenic species formation was influenced by culture regime, with PO4-riboside produced under nutrient rich conditions, whereas Dimethylarsenoacetate (DMAA) was found in old (>42 days old) batch cultures, with this arsenic species possibly produced by the degradation of arsenoribosides-arsenolipids from decomposing cells rather than by biosynthesis. Nutrient availability, hence culture regime was thus influential in directly and indirectly influencing arsenic cycling and the arsenic species produced by D. tertiolecta and T. pseudonana. Future research should thus utilise continuous culture regimes to study arsenic cycling as these are far more analogous to environmental processes.