The aim of the study is to examine the suitability of the Lang II Random Dot Stereotest for use by district nurses in primary healthcare centres. The Lang Stereotest is an inexpensive method. It is easy to apply, but the child's co-operation is required. Approximately 100 district nurses were given detailed information and instruction by an orthoptist prior to a randomized population-based community trial initiated in 1993. In the county of Uppsala in Sweden, all 4293 children born in 1990 were randomized to one of the following four study groups: intervention, control, reliability or validity. All children were followed from birth to 6.5 y of age with respect to the occurrence of strabismus. The kappa statistic for test-retest reliability among district nurses was 0.65 (95% CI, 0.51-0.80). A validation of the outcomes of district nurses' against orthoptists' examinations using the Lang Stereotest showed a sensitivity of 0.33 (95% CI, 0.09-0.65) and a specificity of 0.85 (0.80-0.89). Of 113 cases of strabismus, 53% were detected within the first 3 y of life. Examination of 1132 3-y-old children using the Lang Stereotest enabled the district nurses to detect 11 new cases of strabismus. Only one of these cases was referred to an orthoptist as a direct consequence of the screening. The Lang Stereotest showed low reliability and low validity in this community trial. It is concluded that the Lang II Random Dot Stereotest is not suitable for the screening of manifest strabismus in unselected populations of 3-y-old children in Sweden, where a large proportion of cases of strabismus is detected during the first 3 y of life.