The Wagner equation describing the plasma concentration‐time relationship for the two‐step infusion method is presented in general form for application to both the fast and slow infusion phases of single‐ and multiple‐cycle infusions. The generalized parameters C (steady‐state concentration), Q (infusion rate), A (amount of drug), and t' (time) are defined for each case. Plasma concentrations are simulated for a hypothetical two‐compartment drug, X, to illustrate the uses of the equation and to examine error effects in the QF/QS ratio, T, and ke1. For the single‐cycle case, it is shown that positive or negative (±) errors in QF/QS and T delay the approach to the steady‐state plasma value, Css1, but such errors would not be of clinical importance when they are ⩽ ±2.5%. Errors in Css1, and more greatly affect the plasma concentration at the end of the dose cycle, Css1, than do the same errors in QF/QS or T. Therefore, the use of an average population value for ke1 instead of the actual patient value is more likely to produce large deviations from Css1 than experimental errors in flow rate and T. The utility of the two‐step infusion method in reaching successively higher steady‐state values, Css1,(n), for drugs with a high therapeutic index is also demonstrated for drug X over a range of multiple cycles. The examination of error effects for the multiple‐cycle case revealed that errors in QF/QS and T yield Css1,(n) values that tend to converge toward the desired value of Css1,(n), as the dose cycle increases, while such errors in ke1 yield Css1,(n) values that increasingly diverge from Css1,(n), as dosing increases. Thus, individual patient differences as reflected in ke1 become increasingly more important as a source of error in Css1,(n) with successively higher dosing, while experimental errors in QF/QS become less important. Copyright © 1978 Wiley‐Liss, Inc., A Wiley Company