Although the nonstress test has generally supplanted the contraction stress test as the primary test for assessment of fetal well-being, the contraction stress test continues to be useful. Objectionable factors of the contraction stress test have included the time required to complete the test and the frequent necessity for intravenous oxytocin. The authors developed a contraction stress test protocol using contractions produced by intermittent nipple stimulation and used the method for 345 tests on 193 high-risk pregnancies during a 15-month period. The women were instructed to stimulate one breast, through their clothes, for two minutes and then to rest for five minutes. This cycle was repeated as necessary, but was interrupted whenever contractions began. Using this protocol, the authors were successful in achieving an adequate contraction frequency with every attempt, and hyperstimulation was not more frequent than previously reported with the standard contraction stress test. The average time requirement was 45 minutes. No unexplained stillbirths occurred after a negative test.