The pharmacokinetics of intraperitoneal (i.p.) radiolabelled monoclonal antibody (MAb) was studied in 35 patients receiving 40 i.p. injections. Eleven patients received 131I‐labelled MAb, 24 received 90Y‐labelled MAb, and 5 patients received a second 131I MAb treatment after having developed human anti‐mouse antibodies (HAMA). All patients had blood and urine isotope activity monitored for 5 days after MAb injection. The radiation dose to bone marrow from the vascular compartment in the marrow was calculated by applying the MIRD formula to the measured blood activity. In HAMA negative patients, peak blood isotope activity was observed at 40 hr post injection with a mean of 26% and 21% of the injected 131I and 90Y activity respectively. Sixty‐five percent of the injected 131I activity, but only 12% of the administered 90Y, was excreted in the urine. Myelosuppression limited the administered 131I and 90Y activities to below 160 and 20 mCi respectively. In patients receiving 131I labelled MAbs, the marrow is irradiated by MAb within its circulation, producing myelosuppression that can be predicted by applying the MIRD formula to the blood isotope activity. This is not true for 90Y‐labelled MAbs, where bone absorption of yttrium (which cannot be measured in patients) is the dominant radiation source for bone‐marrow irradiation. Patients with HAMA present clear 131 MAb rapidly with a decreased radiation dose to marrow and reduced myelosuppression. Giving patients intravenous antimouse immunoglobulin to clear 131I‐labelled MAb absorbed from the peritoneal cavity could decrease the toxicity observed in these patients. Patients receiving 90Y DTPA‐chelated MAbs are unlikely to benefit, as catabolized yttrium is not excreted, and is concentrated in liver, spleen and bone. On the other hand, the use of i.v. chelating agents such as EDTA may scavenge non‐protein‐bound (90Y with increased excretion in the urine and less myelosuppression.