Consideration of the clinical results reported, when a cyclotron produced neutron beam was used for treatments in the pelvis region, suggested that a constant RBE of 3 should not have been used for all neutron doses. Instead a variable RBE, which increased from approximately 3 to 8 (with decreasing dose), should have been used. Although some of these RBE values are much higher than 3, they have been observed in clinical practice. An "equivalent photon" isodose plan was produced by employing a variable RBE and, by taking a TDF limit of 86 for bowel, an isoeffect plan was produced. This shows that in the clinical situation under consideration much of the pelvis was overdosed. Doses to tumour cells and late effects are also briefly considered. It is suggested that, in neutron therapy, both an "equivalent photon" isodose plan and an isoeffect plan should be produced prior to treatment.