A cooperative experiment involving 999 litters was conducted at seven stations to assess the reproductive performance of gilts and sows fed different levels of energy during gestation and lactation (21 d) for three consecutive parities. Treatments consisted of two gestation (G) and two lactation (L) diets in a 2 x 2 factorial arrangement. Gestation treatments were diets fed to provide normal (5.9 Mcal/d, NG) or high (7.4 Mcal/d, HG) energy intake; lactation treatments were diets that contained 0 (NL) or 9% added fat (HL). Sows were provided ad libitum access to feed during the lactation period. Feeding the higher energy level during gestation increased pig birth weight (P < .01) and pig weight gain to weaning (P < .01). Sows that had been fed the higher level of energy during gestation (HG) ate less feed during lactation (P < .01) and lost more weight during lactation (P < .01). Increasing lactation energy intake by adding fat resulted in greater pig weight gains to 21 d of age (P < .01). Sows fed the HG-NL combination had fewer pigs on d 21 than all other treatment groups (G x L; P < .01). There was a G x L x parity interaction (P < .05) for days to estrus after weaning. Sows on the NG-HL or HG-NL treatments had increased days to estrus in Parity 1 (P < .05), whereas those on the NG-NL treatments had increased days to estrus in Parities 2 and 3 (P < .05). The HG-HL treatment minimized days to postweaning estrus during the three-parity study.(ABSTRACT TRUNCATED AT 250 WORDS)