Among recent studies considering the splitting of an order among several suppliers (i.e., in “multiple sourcing”), one group considered only the favorable effect of multiple suppliers on the effective lead time demand and required safety stock, but its effects on the annual order and holding cost components have been ignored. Another group that considers the effect of using two suppliers on all relevant cost components imposes severe restrictions on the suppliers' lead-time distributions as well as the proportion of order-split between the suppliers. Our primary purpose is to present easily-solvable decision models for minimizing the sum of annual holding and ordering costs with two suppliers, subject to a maximum allowable stockout risk; restrictions on lead-time distributions and order-split proportion are completely eliminated. Solving these models gives the optimal total order quantity, reorder point and proportion of split between the two suppliers. Numerical results from our models reveal some unexpected observations; e.g.: (i) In using two suppliers, die reduction of inventory carrying cost (a hitherto unrecognized component) is at least as important, and often considerably more important than the effect of safety-stock-cost reduction, (ii) Although intuitively one might use suppliers with the shortest (mean) lead times, it is actually better to have two suppliers such mat the second supplier's mean lead time is “suitably” larger man the first's; this could mean excluding the one with lowest mean lead times among the candidates for the second supplier, (iii) The optimal proportion of split varies with, among other factors, the difference in the suppliers' mean lead times.