Few tuberculosis studies carried out in Kenya since 1948 have reported on the variations of the occurrence and geographical distributions of non-tuberculous Mycobacteria infection. They have however not been able to clearly relate infection outcome to the use of tuberculin tests as epidemiologic and clinical tool. The present survey, conducted by Kenya Medical Research Institute in collaboration with the International Union Against Tuberculosis and Lung Diseases between 1986 and 1990 among school children aged 6-13 years in 18 randomly selected districts, investigated the following: the prevalence of sensitivity to tuberculin PPD RT 23 and PPD scrofulaceum RS 95 sensitin; geographic distribution of the reactions from the tests and; the influence of non-tuberculous Mycobacteria on the tuberculin results. The WHO cluster sampling procedures were used to select 30 schools in each of the 18 districts. Each child was tested with an intradermal dose of 2 TU of PPD RT 23 with Tween 80 (PPD RT 23) on the right hand and 2 TU of PPD Scrofulaceum RS 95 sensitin with tween 80 (PPD RS 95) on the left hand. The results were read after 72 hours of testing. A total of 1015 BCG scar negative children was included in the survey. Of these, 981 were tested and read. Over 47% of the children did not react to both antigens while 6.1% and 22.7% reacted to human type tuberculin and environmental sensitin respectively. There was cross reaction between the two tests in 23.8% of the children from low altitude area who had more and larger reaction to the tests than the middle and higher altitude regions of the country.(ABSTRACT TRUNCATED AT 250 WORDS)