It has been argued that a recently developed measure, vagal tone (V-triple-overdot), is a significant advancement over other existing methods of assessing the periodic variation in heart rate associated with respiration (respiratory sinus arrhythmia). It has been further suggested that, as a noninvasive measure of vagal nerve efferent activity, V-triple-overdot may facilitate the early identification of infants at risk for developmental disabilities. This study addressed the relationship between V-triple-overdot and other measures of cardiac activity and behavioral state and the stability of V-triple-overdot across repeated measures. Twelve samples of cardiac activity were collected from each of 20 full term infants, 6 samples on each of two consecutive days. V-triple-overdot values were derived using a spectral analysis program comparable to Porges' patented MXedit process. Measures of behavioral states were collected by continuous observation. Heart period and heart period variability were highly correlated with V-triple-overdot. Variation in V-triple-overdot between behavioral states was also detected. Repeated assessments revealed that average V-triple-overdot values collected in the same state were not significantly correlated across successive days. This short-term variability both between and within individuals does not support the notion that a single assessment of V-triple-overdot can, by itself, be used to identify at-risk infants or predict developmental outcome.