We measured bone speed of sound in premature infants by quantitative ultrasound. A total of 44 neonates participated in the study including 29 premature infants (median birth weight 1264 g, range 578-2420 g; median gestational age 31 weeks, range 24-36 weeks) and 15 full-term infants (median birth weight 3360 g, range 2700-3730 g; median gestational age 40 weeks, range 37-41 weeks). The left tibial speed of sound (SOS) was measured by quantitative ultrasound. Bone SOS was successfully measured in all infants. We found a significant correlation between tibial SOS and gestational age (r = 0.78, P < 0.0005), but a significant inverse correlation between tibial SOS and post-natal age (r = -0.78, P < 0.0005). Bone SOS was significantly (P < 0.05) higher in full-term infants (3101 m/s, range 2899-3314 m/s) compared to premature infants (2821 m/s, range 2516-3139 m/s), and compared to a subgroup of the premature infants who reached corrected age of full-term infants (2706 m/s, range 2516-2892 m/s, n = 13). Bone SOS was lower (2745 m/s, range 2533-3036 m/s, n = 16) in very low birth weight premature infants (birth weight < 1500 g). CONCLUSIONS The results indicate that tibial speed of sound was reduced in premature infants (in particular very low birth weight) compared to full-term infants even when premature infants reached the corrected age of their full-term peers. The potential role of this technique in assessing osteopenia in premature infants warrants further exploration.