The purpose of this study was to evaluate the validity and reliability of the Futrex 5000A near infrared (NIR) spectrophotometer for the assessment of body composition in children and adolescents. Forty-eight subjects (24 boys and 24 girls) with a mean age of 12.7 +/- 2.7 yr underwent three methods of body composition testing: NIR, densitometry by hydrostatic weighing (HYDRO), and skinfold anthropometry (SKF). Percent body fat (%BF) and fat free mass (FFM) derived from the Lohman's age-adjusted Siri equation served as the criterion. Within session test-retest reliability was determined for the NIR device for all subjects and between-week reliability was evaluated for all test methods in 14 subjects. Based on the excellent within- and between-session reliability (ICC ranged from 0.907-0.999), the system offers the potential of obtaining longitudinal data in growth and development studies. However, the significant mean differences, moderate correlations with the criterion (r = 0.62-0.71) and inflated standard errors of estimate (SEE = 4.9-5.5% BF, 2.2-2.9 kg FFM) and total prediction errors (TE = 5.5-8.0% BF, 2.7-3.7 kg FFM) indicate that refinement of prediction equations is needed to establish the measurement validity. Continued research with expanded populations is needed to further demonstrate and evaluate the utility of this device.