A new surveillance system was initiated on selected growth and nutritional characteristics of children living in inner-city areas and children from ethnic minorities. The heights of Caucasian, Afro-Caribbean and Indo-Pakistani children in this study were compared with those of children in an existing surveillance study, who were chosen to be representative of the English population. Data for this representative sample were collected in 1982 and for the ethnic groups in inner city areas in 1983. The analysis included 13,107 boys and girls aged 5-11 years. Very large differences in height were detected between ethnic groups. The Afro-Caribbean children were the tallest, on average around 3.5 cm taller than the 1982 sample, while the Gujarati children were the shortest, on average about 3 cm below the 1982 sample. Adjustment for a large set of biological and social variables did not eliminate differences in height between ethnic groups. This would indicate that the use of British standards of height based on Caucasian children to assess growth of a child of another ethnic group in England should be interpreted with caution. Multiple regression analyses by ethnic group revealed differences in the pattern of associations between height and social and biological factors among groups. Generalizations from findings in one ethnic group to another in England are not appropriate.