Three flea beetle species (Aphthona spp.), first introduced into North America in 1988, have come to be regarded as effective biological control organisms for leafy spurge (Euphorbia esula). The black flea beetles (Aphthona lacertosa and A. czwalinae) in particular have been shown to cause reductions in leafy spurge stem counts in the northern Great Plains, while the brown flea beetle (A. nigriscutis) has persisted and spread, but has not been found to be as effective at controlling leafy spurge. The ability of black flea beetles to control leafy spurge in any given year, however, has been found to vary. To better understand the long-term effects of flea beetle herbivory on leafy spurge, we monitored stem counts of leafy spurge and numbers of black and brown flea beetles at three sites on two National Wildlife Refuges in east-central North Dakota, USA, from 1998 to 2006. Brown flea beetle numbers were observed to be negligible on these sites. Over the 9 years of the study, black flea beetles were seen to spread over the three study sites and leafy spurge stem counts declined substantially on two of the three sites. Even at low densities of spurge, black flea beetle populations persisted, a necessary prerequisite for long-term control. We used structural equation models (SEM) to assess the yearly effects of black flea beetles, soil texture, and refuge site on leafy spurge stem counts over this time period. We then used equations developed from the SEM analysis to explore flea beetle–leafy spurge dynamics over time, after controlling for soil texture and refuge. Yearly effect strength of black flea beetles on leafy spurge was found to be modest, largely owing to substantial spatial variability in control. However, simulation results based on prediction coefficients revealed leafy spurge to be highly responsive to increases in flea beetle populations on average.