BACKGROUND: The normal clinical index and the standard error of ultrasonic bone density in the population of normalchildren aged 6-12 years have not yet been established. OBJECTIVE: To investigate the status of ultrasonic bone density in the population of normal children aged 6 to 12 years old in Shenzhen City and establish a normal reference index of ultrasonic bone density of this population. DESIGN: Cross-sectional survey. SETTING: Ultrasound Department of People's Hospital in Longgang District of Shenzhen City. PARTICIPANTS: The volunteers for ultrasonic bone density detection were chosen froma kindergarten, an elementary school and a middle school between March 2002 and April2003. Excluded were those who had bone fracture and received medication of hormone and other drugs within 6months that affected bone metabolism, and menstrual history. Altogether we chose 697 normal children, 367 boys and 330 girls whose age ranged from 6 to 12 years old. METHODS: SAHARA ultrasonic bone density apparatus (the US) was used, and corrected according to the standard body model after the operation began.The error of precision was below 1％ and the error of accuracy was 3％. Bone density of the left heel of the children was detected and body mass (kg) and height (cm) were also measured using the same height and weight scale. MAIN OUTCOME MEASURES: The correlation between bone densityof the heel and sex, age and body mass in the population of normal children aged 6 to 12 years. RESULTS: The normal reference values of bone density in the population from 6 to 12 years old were (0.445±0.166) g/cm2 in 6-year-old group,(0.509±0.151) g/cm2 in 7-year-old group, (0.510±0.133) g/cm2 in 8-year-old group, (0.519±0.132) g/cm2 in 9-year-old group, (0.520±0.153) g/cm2 in 10-year-old group, (0.53±0.175) g/cm2 in 11-year-old, and (0.545±0.206) g/cm2 in 12-year-old group. There were differences in bone density between boys and girls (P ＜ 0.001), and after correction by body mass, the difference disappeared. Bone density of the boys and girls had increasing linear correlation with age (male r=0.722, P ＜ 0.001; female r=0.785, P ＜ 0.001), and had significant relationship with body mass (r=0.984, P ＜ 0.001). CONCLUSION: In normal children aged 6 to 12 years old, bone density of the heel has no association with sex, but has increasing linear correlation with age and significant relationship with body mass.