Estimating the effect of trade on capital flows is difficult given the inherent identification problem. We use fluctuations in rainfall to capture the exogenous variation in trade between Germany, France, the U.K., and the Ottoman Empire during 1859-1913. The provisionistic policy of the Ottoman Empire--only surplus production was exported--constitutes the basis of our identification strategy. We find that one standard deviation in rainfalls from the mean leads to a 3.5 percent increase in Ottoman exports, which in turn causes a 10 percent increase in capital inflows from the three source countries. Our findings support trade theories predicting complementarity between trade and capital flows.