Background: The fetal origins hypothesis poses that adverse intrauterine conditions predispose to cardiovascular and metabolic diseases in adulthood. Evidence is accumulating that similar mechanisms to those identified for physical disorders may also apply to psychiatric disorders. Focusing on the activity of neurophysiological systems thought to regulate emotions from very early in life may be key to understanding how maternal stress in pregnancy impacts on the developing baby with possible long-lasting consequences for behaviour and psychopathology. Respiratory Sinus Arrhythmia (RSA), "vagal tone", is thought to reflect autonomic regulatory capabilities that may underpin emotion regulation. However, little is known about possible fetal origins of vagal tone. Animal studies increasingly point to sex differences in the effects of prenatal stress, and this is supported by human studies of the prenatal origins of cardiovascular functioning and psychopathology. The current investigation examines whether prenatal depression and anxiety predict vagal tone in infancy, and whether the associations are modified by infant sex. Method: Two hundred mothers and infants from a high-risk consecutive community sample were examined prospectively from the first trimester of pregnancy until 29 weeks postnatal. Maternal self-reports of stress (EPDS and STAI) were collected in pregnancy (20 and 32 weeks) and postnatally (5 weeks and 29 weeks). Vagal tone was ascertained across five procedures, the "Helper-Hinderer" social evaluation task, toy exploration and the "Still Face" paradigm (2 minutes of social engagement, followed by 2 minutes of maternal unresponsiveness and concluded by 2 minutes of social reunion). Results: Principal Component Analysis of the RSA scores yielded a one-factor solution explaining over 70% of the variance, and so mean of RSA scores was used as the index of overall vagal tone, and the difference between overall and RSA during the Still Face as the estimate of vagal withdrawal. There were no main effects of prenatal maternal depression or anxiety on vagal tone or vagal withdrawal. However, there were significant prenatal stress by sex of infant interactions. Follow-up analyses revealed that increasing maternal depression and anxiety at 20 weeks gestation were associated with decreasing vagal tone in males and increasing vagal tone in females. Vagal withdrawal in response to the still face showed similar patterns i.e. decreased in males and increased in girls with elevated maternal anxiety at 32 weeks gestation. These associations were not explained by possible confounding variables assessed in pregnancy, nor by postnatal maternal depression and anxiety. Conclusions: The findings support the fetal origins hypothesis for vagal tone and vagal withdrawal, but only in interaction with sex of the infant. Longitudinal study is required to determine conditions under which increasing vagal tone and withdrawal in girls associated with prenatal depression and anxiety, and decreasing vagal tone and withdrawal in boys, are associated with later resilience or vulnerability to psychopathology.