Observed and disease-free survivals were evaluated in a consecutive series of 46 resections of pulmonary metastases, with major chance of being a unique phenomenon. Survival curves were computed both since the treatment of primary tumors and since resection of lung metastases. From the treatment of primary tumors, median disease-free interval was 33 months, and rose to 66 months after resection of lung metastases. From the treatment of secondary lung cancers the observed survivals at 1, 3 and 5 years were respectively 60%, 41% and 26%. Survival was clearly affected by development and resectability of post-thoracotomic recurrence (100% without recurrence, 50% with resectable recurrence and 4% with unresectable recurrence). Recurrence rate was related to the first disease-free interval and to the anatomical extent (particularly for nodal status) of secondary lung cancer. This fact suggests that the failure of secondary lung cancer resection may arise either from the primary cancer (poor selection) or from the secondary cancer (delay in treatment).