Rubella is a systemic virus infection that is usually mild. It can, however, cause severe birth defects known as the congenital rubella syndrome (CRS) when infection occurs early in pregnancy. As many as 8%–13% of children with CRS developed autism during the rubella epidemic of the 1960s compared to the background rate of about 1 new case per 5000 children. Rubella infection and CRS are now rare in the U.S. and in Europe due to widespread vaccination. However, autism rates have risen dramatically in recent decades to about 3% of children today, with many cases appearing after a period of normal development (‘regressive autism’). Evidence is reviewed here suggesting that the signs and symptoms of rubella may be due to alterations in the hepatic metabolism of vitamin A (retinoids), precipitated by the acute phase of the infection. The infection causes mild liver dysfunction and the spillage of stored vitamin A compounds into the circulation, resulting in an endogenous form of hypervitaminosis A. Given that vitamin A is a known teratogen, it is suggested that rubella infection occurring in the early weeks of pregnancy causes CRS through maternal liver dysfunction and exposure of the developing fetus to excessive vitamin A. On this view, the multiple manifestations of CRS and associated autism represent endogenous forms of hypervitaminosis A. It is further proposed that regressive autism results primarily from post-natal influences of a liver-damaging nature and exposure to excess vitamin A, inducing CRS-like features as a function of vitamin A toxicity, but without the associated dysmorphogenesis. A number of environmental factors are discussed that may plausibly be candidates for this role, and suggestions are offered for testing the model. The model also suggests a number of measures that may be effective both in reducing the risk of fetal CRS in women who acquire rubella in their first trimester and in reversing or minimizing regressive autism among children in whom the diagnosis is suspected or confirmed.