Spinal posture assessment is an essential component in physical evaluation. Establishing reproducible value of spinal curves is necessary for detecting postural changes. Therefore, the main objectives of this study were to assess the reproducibility of spinal curves in relaxed standing of healthy subjects and to explore the extent and pattern of change upon transition from relaxed to a fully erect posture. Thirty young women and men were measured twice over a 1 week interval for recording the cervical, thoracic, and lumbar curves values using an ultrasound-based system. Thereafter, additional 28 men and women extracted from the same reference group were assigned for a single measurement session, in which the same angular values were obtained in relaxed and fully erect standing postures, as well as stature using stadiometer. Excellent, good, and poor reproducibility indices, standard error of measurement (SEM), and smallest real difference (SRD), were noted for the thoracic (interclass correlation coefficient, ICC(3,3)=0.95, SEM=1.2°, SRD=3.3°), lumbar (ICC(3,3)=0.85, SEM=2.6°, SRD=7.2°), and cervical (ICC(3,3)=0.68, SEM=3.8°, SRD=10.5°) curves, respectively. Erecting from relaxed posture was associated with a significant (>SRD) thoracic angle difference of 7.2° in men and 4.8 in women which was expressed in a height increase of 1.3 cm in men and 0.8 cm in women. These changes were significantly larger in men in whom the angular and height differences were also significantly correlated (r=0.7). Using this system, angular measurements are highly reproducible in the thoracic curvature in young healthy adults, leading to the use of the associated SRD as a criterion for thoracic postural flexibility.