Although it is often assumed that fetal heart rate variability reflects simple "push-pull" interactions between the parasympathetic and beta-sympathetic limbs of the autonomic nervous system, there has been little direct experimental evidence to support this view. We used autonomic blocking agents to investigate heart rate variability in chronically catheterized fetal lambs, and an on-line computer to make measurements of heart rate and of the higher and lower frequency components of its variability. beta-sympathetic blockade alone had no effect on variability. Parasympathetic blockade alone reduced it, but did not abolish it. Even after double blockade, some 35% to 40% of variability remained, thus implying that there is a major nonneural component to heart rate variability. There was evidence that the fetus, unlike the adult, is subject to a resting cardioacceleratory drive.