The effects of gestation housing systems on sow and litter performance were evaluated for 2.5 yr in southwest Iowa. Gestation housing system treatments were as follows: 1) individual gestation stalls in a mechanically ventilated confinement building with a partially slatted floor and a manure flush system and 2) group pens with individual feed stalls in deep-bedded, naturally ventilated hoop barns. In all, 957 litters from 353 sows were evaluated. Number of pigs born alive per litter differed for the 2 housing treatments (P = 0.002). Sows gestated in hoop barns gave birth to more live pigs per litter (10.0 +/- 0.2 pigs) than sows gestated in stalls (9.3 +/- 0.2 pigs). Preweaning mortality was not different for the 2 housing treatments (P = 0.70). Cross-fostering was done to equalize litter size within 24 h of birth, which resulted in an equal number of weaned pigs per sow (P = 0.50) regardless of gestation housing treatment. The weaning-to-breeding interval was different (P = 0.01), with sows kept in stalls (4.3 +/- 0.6 d) returning to estrus sooner than sows gestated in hoop barns (6.0 +/- 0.6 d). These results indicate that gestating sows can be housed as groups in deep-bedded hoop barns equipped with individual feeding stalls and will perform comparably to gestating sows housed in confinement systems with individual gestation stalls.