Addition of ethanol ( ET) to the drinking fluid of pregnant rats has been questioned as an experimental model for the fetal alcohol syndrome (FAS). This model, however. closely simulates human alcohol intake, and in this study we used a modified version of previous protocols to overcome their major defects. A group of female rats was given I 07r ET in drinking fluid for one week, 15'/r for the second week. 20'/r for the third, and 251/r for the fourth, at the end of which they were mated with non-treated males and given 251/r ET throughout gestation. Three groups of non-ET treated sex and age-matched rats were studied in parallel: (I) normal controls receiving solid diet ad lib. (2) paired fed rnts, and (3) rats fed ad lib the solid diet mixed with 50'/r fiber. In the ET group, food intake decreased as ET consumption augmented, the ET calories comprising over 3(flr of the total energy intake during pregnancy. Total energy intake was similar for ET group and normal controls, and was higher than in paired fed animals or those on 5(f/r fiber diet. Body weight gain in ET rats was similar to those on 50'/r fiber diet, lower than in normal controls and higher than in paired fed animals. At the 21st day of gestation. rats on ET had plasma ethanol levels of 147:+: 18 mg/di and higher plasma osmolality than in the other groups studied. In ET rats, fetal body weight was lower than in either normal controls or rats on 5(f/r fiber diet, and fetal body length was shorter than in any other group. These findings demonstrate that our protocol provides a suitable animal model for the study of FAS, and indicate that rats on 507r fiber diet are better control subjects than paired fed rats.