Integrated pest management (IPM) for invasive plant species is being advocated by researchers and implemented by land managers, but few studies have evaluated the success of IPM programs in natural areas. We assessed the relative effects of components of an IPM program for leafy spurge (Euphorbia esula), an invasive plant, at Theodore Roosevelt National Park, North Dakota. Effects of herbicides on leafy spurge abundance and on dynamics of flea beetles (Aphthona spp.) used to control leafy spurge were evaluated over three field seasons following herbicide application. We monitored leafy spurge-infested plots with established flea beetle populations that had received picloram plus 2,4-D in September 1997 or 1998, imazapic in September 1998, versus those with no chemical treatment. Mature stem counts did not differ significantly between treated and untreated plots in 2001, suggesting that leafy spurge stands had recovered from herbicide treatment. Flea beetles were less abundant on plots with a history of herbicide treatment. Structural equation models indicated that in 2000 negative correlations between relative abundances of the two flea beetle species were greater on plots that had received herbicide treatments than on those that had not, but by 2001 no differences were apparent between treated and untreated plots. These results suggest that the most effective component of IPM for leafy spurge at this site is biological control. All herbicide effects we observed were short-lived, but the increased negative correlation between flea beetle relative abundances during 2000 implies that herbicide application may have temporarily disrupted an effective biological control program at this site. ?? 2006 Elsevier Inc. All rights reserved.