BACKGROUND. Two thrombosis models for the evaluation of thrombolytic agents in small animals (less than 100 g) were evaluated: an iodine-125 fibrin-labeled rat plasma clot in the inferior caval vein of 3-4-week-old rats and a pulmonary embolus in adult hamsters that had been obtained by injection of a 125I fibrin-labeled human plasma clot. The extent of thrombolysis was determined by continuous external monitoring of radioisotope over the thrombus region and by ex vivo recovery of residual clot. METHODS AND RESULTS. In the rat model, infusion of solvent for 60 minutes was associated with mean +/- SEM lysis within 90 minutes of 13 +/- 3% (n = 8) by external counting and 26 +/- 4% (n = 8) by radioisotope recovery. Intravenous infusion of recombinant tissue-type plasminogen activator (rt-PA) over 60 minutes caused dose-dependent progressive clot lysis; with 0.5 mg/kg, producing a plasma level of 0.14 +/- 0.04 microgram/ml, lysis was 64 +/- 9% (n = 4) by external gamma counting and 78 +/- 4% (n = 4) by residual isotope in the vein segment and was not associated with significant fibrinogen or alpha 2-antiplasmin breakdown. In the hamster model, infusion of solvent for 60 minutes was associated with lysis within 90 minutes of 19 +/- 4% (n = 11) by external gamma counting and 31 +/- 3% (n = 14) by residual radioisotope. Intravenous rt-PA during 60 minutes resulted in dose-dependent progressive thrombolysis; with 0.5 mg/kg, producing a plasma level of 0.14 +/- 0.01 micrograms/ml, lysis was 50 +/- 4% (n = 4) by external gamma counting and 78 +/- 5% (n = 4) by residual radioactivity, without an extensive decrease in fibrinogen or alpha 2-antiplasmin. Parallel experiments in the rabbit jugular vein thrombosis model with a rabbit blood clot with intravenous infusion over 4 hours produced 7 +/- 2% (n = 9) lysis with solvent and dose-dependent progressive lysis with rt-PA; with 1 mg/kg, producing a plasma level of 0.20 +/- 0.03 microgram/ml, lysis was 56 +/- 7% (n = 7) by external gamma counting and 61 +/- 7% (n = 7) by residual radioactivity, without extensive consumption of fibrinogen or alpha 2-antiplasmin. CONCLUSIONS. These two thrombosis models in small animals are as reproducible and quantitative as the extensively used rabbit jugular vein thrombosis model. The hamster pulmonary embolism model is superior because it is simpler and more straightforward and allows the performance of as many as 10 experiments by one investigator in 1 day.