We investigate the contagion hypothesis between the United States and three European markets (Germany, the United Kingdom, and France). We focus on realized volatility, which we break down into continuous and jump parts, and we test the contagion hypothesis between jumps during overlapping and non-overlapping hours. We find a significant relation between jumps and realized volatility and spillover effects between jumps. The U.S. market plays the leading role during overlapping hours, but regional contagion is more obvious during non-overlapping hours. Interestingly, jump contagion effects exhibit asymmetry and nonlinearity, and vary according to regimes. Accordingly, we improve jump modeling and spillover.