The primary purpose was to investigate the efficacy of a full year of single-sex (SS) teaching of science. The secondary aims were to locate any differentiation by set and gender, and to relate these to more proximal variables. Participants were 13 year olds. Higher set girls gave evidence of clear benefits overall, and higher set boys also, except in biology. Lower set pupils performed at or below expectations. Analyses of additional questionnaire and interview data pointed to further reasons for avoiding the making of unqualified generalizations about SS teaching. Pupil preferences for SS teaching were relevant, as were gender differences in attitudes to biology and physics. Qualitative data suggested higher set girls benefited from more learningrelated classroom interaction and less interference and exploitation of girls by boys in SS classes. Lower set pupils complained that SS teaching deprived them of social interaction with the other sex. The concluding suggestion was that SS teaching offers affordances of benefits when mixed-sex teaching has specifiable disadvantages.