We study the impacts of in utero exposure to Hurricane Catarina of March 2004, the first hurricane to hit Brazil. Catarina was unexpected and is representative of other recent hurricanes in the Americas in terms of wind speed, direct economic costs, and population affected. We use a triple differences strategy (close vs. far municipality, 2004 vs. 2003, after March vs. before) to highlight the importance of accounting for flexible season of birth effects compared to a standard differences-in-differences strategy. Using administrative data, we find that average birth weight declined and post-neonatal mortality increased among babies exposed to the hurricane in utero. The adverse effects are driven by babies of younger mothers. Our documented impacts are not explained by reductions in employment or healthcare use. Maternal stress seems to be a plausible mechanism if younger women are more financially vulnerable to negative shocks, consistent with recent work highlighting the relationship between socioeconomic status, stress, and birth outcomes. Our findings are robust to various checks, including testing for pre-trends in infant health outcomes.