This study develops and estimates mixture models of crop price comovements using copula functions, which allow for departures from normality during extreme market circumstances. The models also account for unique time-series patterns inherent in crop price data. The results point to two main conclusions. First, mixture models appear to provide an easy-to-estimate approach for capturing real-life crop price movements. Second, mixture models find that, during extreme market downswings, correlations in price movements strengthen by several orders of magnitude. These results suggest that structured securities assembled from different crops tend to lose diversified protection during extreme market downswings, the exact times when such protection is needed most.