Summary.Previous investigations of the effect of (English‐Welsh) bilingualism on performance in non‐verbal tests of intelligence have not yielded a consistent pattern of results. In the present study a non‐verbal test was given to ten‐year‐old pupils from primary schools in Wales. Linguistic background was assessed by means of a language questionnaire, the assessment being compared with teachers' ratings. Care was taken to select a homogeneous group of a thoroughly Welsh background (i.e., a bilingual group) together with a corresponding English‐speaking group (with a background of no appreciable ‘Welshness’). The test results showed a statistically significant difference, in favour of the monoglots, corresponding to about 8 points of I.Q. It is suggested that this difference may be due, in part, to (i) the test being a timed one, and (ii) the groups not being equal with respect to urban‐rural differences. Again the ‘verbal residue’ of a non‐verbal test might well be a non‐negligible factor in measuring the intelligen