The toxicity of arsenate (As(V)), arsenite (As(III)), monomethylarsonic (MMAA) and dimethylarsinic acid (DMAA) to natural algal assemblages from lakes within the Aberjona watershed having different arsenic concentrations was determined by a short-term photosynthesis assay. Total arsenic concentrations in the studied lakes ranged from 1.5 x 10(-8) to 1.9 x 10(-7) M. The toxicity of the arsenic species generally decreased in the order of As(V) > As(III) > DMAA for all lakes. Toxicity of As(V) to phytoplankton collected from relatively unpolluted Horn Pond was highest (EC50 = 3 x 10(-7) M), whereas algae from the polluted Upper Mystic Lake were more tolerant (EC50 = 6 x 10(-6) M) and those from Halls Brook Storage Area, a highly contaminated lake, were tolerant to As(V) up to 10(-3) M. The sensitivities of the different algal communities to As(III) were similar (EC50 value = 5 x 10(-5) M). MMAA was as toxic as As(V) in the unpolluted system (Horn Pond), However, photosynthesis of lakes from contaminated lakes was slightly enhanced by MMAA. DMAA concentration in the range of 10(-5) M was strongly enhancing to short-term CO2 fixation rates of all phytoplankton assemblages; activity increased up to 600% compared to control values. Algae from contaminated sites appear to have adapted to higher As(V) and MMAA concentrations, whereas algae from the unpolluted lake remained sensitive. As(III) represented no chemical stress to the algal assemblages as measured by the assays. The role of DMAA is unclear. Considering the concentrations and the toxicity of As(V) and MMAA in unpolluted systems, one must conclude that As(V) is the major arsenical environmental hazard, however, MMAA might also pose a potential risk in unpolluted systems. (C) 1999 Elsevier Science B.V. All rights reserved.