The 'trade promotes peace' hypothesis rests upon three premises: (1) Societies achieve salient economic gains from their trading relationships; (2) serious conflict among societies disrupts trade; and (3) premises 1 and 2 enter the calculus of political decisionmakers. If any one of the three premises does not hold, the liberal linkage between trade and peace is broken. In a recent analysis of seven non-major power dyads, Katherine Barbieri & Jack Levy raise doubts about the second premise upon which the liberal hypothesis rests. In this article, we further test the trade disruption premise. We use an interrupted time-series model to study the impact of war on trade for 14 major power dyads. We find reasonably strong evidence that major power war is associated with a decline in trade relative to pre- and postwar periods. We also investigate the impact of war on trade for 13 non-major power dyads. Here the evidence is weaker but on balance remains supportive of the trade disruption premise.