An analysis is presented of a double-junction technique for the treatment of the complete CNS using orthogonal brain and spine ports from a 4-MV linear accelerator with a lead flattening filter. Experimental results are compared with calculated results in the junction zone. Clinically homogeneous doses (.+-. 6%) can apparently be delivered to the total CNS, provided that the irradiation of the junction zone is terminated .apprx. 10% sooner than irradiation of the rest of the brain and cord. The number of reduced fractions to be delivered in this region can be determined from the results of scatter dose calculations.