Tumor therapy using radiolabeled antibodies presents a challenging problem in absorbed dose determination. The purpose of this study is to evaluate the effect of tumor size on the absorbed dose distribution from beta‐emitters when the radiolabeled antibody is not uniformly distributed throughout the tumor. Two theoretical dosimetry models are constructed, one for nonvascularized micrometastases and the other for vascularized tumors. All calculations assume no penetration of radionuclide into the tumor. These are compared to an even distribution of radionuclide throughout the tumor. In micrometastases of 1‐mm diameter or less, emitters of low energy such as 131I give higher dose rates than emitters of higher energy because less energy is lost outside the target volume. However, even with 131I, a significant proportion of the energy is not absorbed in the tumor and, as a result, the concentration of radionuclide necessary for a therapeutic radiation dose becomes higher as the tumor diameter gets smaller. Because it may be impossible to achieve these concentrations in very small tumors (<0.5‐mm diameter), alpha‐emitters may be useful in combination with beta‐emitters for therapy of micrometastatic disease. In vascularized tumors, higher energy emitters such as 90Y yield higher doses because of overlapping dose distributions from multiple vascular sources. This also produces a more even dose distribution across a tumor, even when there is poor penetration of the radiolabeled antibody. Thus tumor size, antibody penetration, and tumor vascularity all influence the choice of radionuclide and, depending on the circumstances, alpha‐emitters, low‐energy beta‐emitters, high‐energy beta‐emitters, or some combination of the three may be most efficacious.