A proposed mechanism [Clin. Chem. 19, 668 (1973)] for the inverse relationship beteen guanidinoacetate (I) and guanidinosuccinate (II) in human urine is explored. The mechanism proposes that canavaninosuccinate (III) may be reduced to form homoserine and II or, alternatively, that the III may be acted upon by a lyase to form canavine and fumarate. The canavanine would the proceed to transamidinate to glycine to form I. This study demonstrates for the first time that layse activity for converting III to canavanine and fumarate exists in human liver and kidney extracts. Transamidination from canavine to glycine to form I is also readily accomplised with human tissue. Reductive cleavage of III to II and homoserine has been demonstrated before [Clin. Chem. 15, 397 (1969)]. The optimum pH for the lyase reaction is 6.5, for the reductive cleavage it is 8.7. In follwing the course of the lyase reaction, we developed a technique whereby the fumarate formed was hydrated with fumarase (EC 22.214.171.124) and then dehydrogenated with malate dehydrogenase (EC. 126.96.36.199). The changes in absorbance of NADH formed in the reaction were then measured and used to determine the amount of fumarate formed, as a measure of lyase activity. Canavanino-succinate lyase activity follows pseudo-first-order reaction kinetics. The Michaelis constant of this lyase was 6.16 X 10-4 mol/liter, for argininosuccinate lyase 9.74 X 10.4 mol/liter. These data suggest that the binding afficinity for III to the enzyme is greater than that for argininosuccinate. Glycine added to the reaction acts as an activator, probably because is removes the canavanine from the reaction mixture. On the otherhand, arginine acts as an inhibitor of III-lyase. Other substances tested, such as canavaine, fumarate, and arginissuccinate had no effect on the reaction kinetics.