Two mutations of purine nucleoside phosphorylase (PNP) were identified in the carrier state of the first generation progeny of male mice treated with ethylnitrosourea mated to untreated females. The variants are assigned the gene symbols Np-1e and Np-1f. Both carriers have approximately half normal PNP activity in erythrocytes and Np-1a/Np-1a was distinct from the inbred strain background, Np-1a/Np-1a on isoelectric focusing, whereas Np-1a/Np-1a was not distinguishable. In the homozygous state, Np-1a and Np-1a differ and determine a more basic pattern of PNP activity than the Np-1a allele. NP-1A is stable in the presence of phosphate at 55° C whereas the half lives for the mutants were 30 and 7.5 min for NP-1E and NP-1F respectively. The substrate Michaelis constants for the variants were unchanged from controls and the maximal velocities for erythrocytes were: 16.8 ± 1.1, NP-1A; 2.16 ± 0.12, NP-1E; and 0.50 ± 0.03, NP-1F (nmole/min/mg protein). Brain, heart, kidney, liver, spleen leukocytes and thymocytes also showed reductions in activity, 4-27% of normal for NP-1E and 0.1-3.9% for NP-1F. Purine nucleoside excretion correlated with the severity of the enzyme deficiency. The substrates of PNP are not found normally ingurine < 10 uM, but inosine and guanosine were present for Np-1e/Np-1e mice, total 150 ± 84 uM; as were inosine, guanosine, deoxyinosine, and deoxyguanosine for Np-1f/Np-1f mice, total 1490 ± 190 uM. Supported by the Medical Research Council grant MT-6376.