Background. Both working memory capacity and cognitive style have independently been found to affect performance on school-type tasks, but their effects in interaction have not been considered. Aims. The aims of this study were to consider the relationship between working memory, cognitive style and gender on (a) overall learning behaviour, and (b) performance on a range of school subjects. Sample. The sample comprised 206 13-year-old secondary comprehensive school Year 8 pupils, being all pupils in that year who completed the principal assessments. Method. The pupils did an assessment of working memory efficiency - the Information Processing Index (Riding, 2000a). They also completed the Cognitive Styles Analysis (Riding, 1991) to determine their positions on the two fundamental cognitive styles, which were indicated by two ratios: the Wholist-Analytic ratio and the Verbal-Imagery ratio. Overall learning behaviour was rated by the pupils' tutors. In addition, attainment in each of 10 subjects was rated by their subject teachers. Results. For overall learning behaviour, there was an interaction between working memory capacity and cognitive style. With the Wholist-Analytic style dimension, memory made a marked difference for Analytics but had little effect for Wholists, and with the Verbal-Imagery dimension Verbalisers were affected but not Imagers. With the school subjects, these differed in terms of their sensitivity to gender, memory and style. Conclusion. The results were discussed in terms of differences between the styles in terms of information-processing demands. Practical ways of improving learning performance were also considered.