Resting growth hormone and prolactin levels and dynamic responses to bromocriptine and metoclopramide have been measured in epileptic patients before treatment, and compared with a matched group taking phenytoin alone. Mean resting levels of prolactin were higher in patients taking phenytoin (untreated patients 204 mU/l, phenytoin treated patients 302 mU/l), but dynamic responses to metoclopramide and bromocriptine were unaffected. Mean resting levels of growth hormone were also higher in patients taking phenytoin (untreated patients 1.4 mU/l, phenytoin treated patients 6.0 mU/l) and paradoxical suppression was seen following bromocriptine. Phenytoin is unlikely to have any major action on the D2 receptor present on the lactotroph. The abnormalities in growth hormone may explain the well recognized effects of phenytoin on connective tissue.