This paper explores the differences between demand and supply uncertainty in multi-echelon supply chains. Our central premise is that the insights gained from the study of one type of uncertainty often do not apply to the other. In fact, the two types of uncertainty are in a sense mirror images of each other, in that the optimal strategy for coping with supply uncertainty may be exactly opposite to that for demand uncertainty. We present several studies, each involving a fundamental question of order frequency, inventory placement, or supply chain structure. In each study, we consider a simple multi-echelon supply chain, first under demand uncertainty and then under supply uncertainty in the form of disruptions. Using simulation, we demonstrate that the optimal strategy is different under the two types of uncertainty and discuss reasons for the differences. We begin with a study demonstrating that the cost of failing to plan for disruptions is greater than that of failing to plan for demand uncertainty. We conclude a with a study that demonstrates that resiliency to disruptions is often inexpensive, in that large improvements in service level can be achieved with only small increases in cost. The remaining studies identify a number of important properties of supply chains subject to disruptions that have not previously been discussed in the literature. For example, we introduce the “risk-diversification effect, ” which describes the benefit of having multiple, decentralized stocking locations under supply uncertainty. This is the opposite of the classical risk-pooling effect, which describes the reverse tendency under demand uncertainty. 1 1.