Correction of a left congenital diaphragmatic hernia in a human fetus with a large volume of liver in the chest requires reduction of liver and viscera into the abdomen. This can kink the ductus venosus and cause the death of the fetus. Therefore, we have repaired surgically created diaphragmatic hernias in fetal lambs by leaving viscera in the chest wrapped in a silastic chimney. With fetal growth there is a relative reduction of hernia volume over weeks, potentially avoiding kinking the ductus venosus. In four groups of lambs lung size and static respiratory system complicance at birth were compared. Lambs treated by this new technique (silo, n = 7) were compared with lambs that had undergone immediate complete correction with a flat silastic patch in the diaphragm plus an abdominal patch (patch, n = 8), with lambs with uncorrected hernias (n = 6), and with normals (n = 8). There was no significant difference between total lung weights (131 ± 6 g v 157 ± 13 g, mean ± SEM, silo v patch) and lung displacement volumes (142 ± 7 mL v 162 ± 14 mL) in either surgically corrected group. Lungs from those corrected by silo were significantly heavier than those with uncorrected herniae (131 ± 6 g v 56 ± 5 g, P < .01), but were not as heavy as normal lungs (131 ± 6 g v 257 ± 16 g, P < .01). Lung displacement volumes in the silo group were significantly larger than those with uncorrected hernias (142 ± 7 mL v 54 ± 7 mL, P < .01), but not as large as normal lambs (142 ± 7 mL v 280 ± 17 mL, P < .01). When respiratory system compliance was standardised for body length (cm) there was no significant difference between the two surgically corrected groups (0.035 mL/cm H 2O/cm ± 0.005 v 0.036 mL/cm H 2O/cm ± 0.004). Compliance was not significantly better than for uncorrected lambs in either of the surgically corrected groups taken on their own, but when the data from both were pooled and reanalysed, the difference between the corrected lambs (n = 15) and those with uncorrected hernias became significant (0.035 mL/cm H 2O/cm ± 0.003 v 0.015 mL/cm H 2O/cm ± 0.002, P < .05). Compliance for normal lambs was significantly greater than for those with a silo (0.062 mL/cm H 2O/cm ± 0.005 v 0.035 mL/cm H 2O/cm ± 0.005, P < .05). When compliance was corrected for lung displacement volume, each unit volume of lung apparently had the same compliance no matter whether from normal, corrected, or uncorrected fetuses.