The release of the seed eating beetle Sulcobruchus subsuturalis for biological control of Caesalpinia decapetala in South Africa has been ongoing since 2000. This is the first post release evaluation of the efficacy of the agent against the weed. The study assessed the phenology of C. decapetala and S. subsuturalis at two sites and included determining the establishment of the beetle on the target weed at 25 study sites. In addition, the effects of S. subsuturalis on seed densities as well as seed germination and seedling recruitment were examined. Furthermore predation by native ants and attacks by native parasitoids were also examined at two release sites. Mature filled pods were available on the tree from September (spring) to March (summer) when the beetle is expected to be reproductively active. The majority of beetle releases by Working for water took place in summer. However the proportion of beetle infested seeds, which were only recovered inside pods in the tree canopy at study sites, was low (0 to 15.5%). Consequently high seed densities and seedling recruitment were observed in the field. In the laboratory, S. subsuturalis did not lay eggs on buried seeds, however seed infestation levels were generally high. Only 8.3% of the seeds containing adults germinated and 6.3% emerged into seedlings. Of the seeds containing larvae, only 14.6% germinated and 2.1% emerged into seedlings. Thus far, S. subsuturalis has failed to maintain high populations on the target weed, possibly due to egg predation by native ants (Crematogaster species, Pheidole megacephala, Messor natalensis and Tetramorium avium) and attacks by native parasitoids. Egg predation reached 100% within 10 days. Egg parasitism by unknown parasitoids ranged between 80 and 93.1% in Limpopo. The parasitic wasp Dinarmus altifrons was recovered once developing inside a C. decapetala seed infested with S. subsuturalis. Further investigations regarding attacks by indigenous parasitoids on S. subsuturalis are required to determine the extent of this problem. Meanwhile releases of S. subsuturalis against C. decapetala should continue. However release efforts need to be improved by (a) using inundative rather than inoculative releases (b) Releasing every year between September and March (summer) and (c) involving land owners, farmers and nature reserve authorities in release activities. Ultimately an additional agent should be sought for release against C. decapetala as the weed may not be the preferred host for S. subsuturalis.