A field trial was conducted for a portable Knee Height Measuring Device (KHMD). Measurement errors for the machine, within and between observers (intraobserver and interobserver) and within subjects (intrasubject), were calculated from repeated observations of two anthropometrists on 40 children, 8 and 9 years of age. Interobserver technical error (0.452 mm) was larger than that published by the KHMD developers, but intraclass reliabilities are still very high (0.9975–0.9983). Attained size and fatness of subjects were not significantly correlated with interobserver measurement errors. Observed errors suggest they are sufficient to allow detection of knee height growth over periods of time much shorter than could be accomplished using more conventional anthropometry.