Intracavitary injections of 32P chromic phosphate are used in the therapy of pleuropulmonary blastoma and pulmonary sarcomas in children. The lung dose, however, has never been calculated despite the potential risk of lung toxicity from treatment. In this work the dosimetry has been calculated in target tissue and lung for pediatric phantoms. Pleural cavities were modeled in the Monte Carlo code MCNP within the pediatric MIRD phantoms. Both the depth-dose curves in the pleural lining and into the lung as well as 3D dose distributions were calculated for either homogeneous or inhomogeneous 32P activity distributions. Dose-volume histograms for the lung tissue and isodose graphs were generated. The results for the 2D depth-dose curve to the pleural lining and tumor around the pleural cavity correspond well with the point kernel model-based recommendations. With a 2 mm thick pleural lining, one-third of the lung parenchyma volume gets a dose more than 30 Gy (V30) for 340 MBq 32P in a 10 year old. This is close to lung tolerance. Younger children will receive a larger dose to the lung when the lung density remains equal to the adult value; the V30 relative lung volume for a 5 year old is 35% at an activity of 256 MBq and for a 1 year old 165 MBq yields a V30 of 43%. At higher densities of the lung tissue V30 stays below 32%. All activities yield a therapeutic dose of at least 225 Gy in the pleural lining. With a more normal pleural lining thickness (0.5 mm instead of 2 mm) the injected activities will have to be reduced by a factor 5 to obtain tolerable lung doses in pediatric patients. Previous dosimetry recommendations for the adult apply well down to lung surface areas of 400 cm2. Monte Carlo dosimetry quantitates the threedimensional dose distribution, providing a better insight into the maximum tolerable activity for this therapy.