Initial field releases of the saltcedar leaf beetle, Diorhabda carinulata Desbrochers, 1870 (Chrysomelidae), against saltcedars, Tamarix Linneaus, 1753 (Tamaricaceae) in North America were unsuccessful at sites where the target taxon was T. parviflora de Condolle, 1828 as opposed to the more widespread T. ramosissima Ledebour, 1829 and related forms. A series of field and greenhouse studies was conducted to determine the basis for these failures. Generally, T. parviflora was a suitable host for larval and adult development. Larval growth was not significantly affected by host species although developmental rate was slightly slower when fed T. parviflora vs. T. ramossisima; nitrogen enrichment had greater influence on growth than did host identity. Insect feeding impact to T. ramossisima outplanted to the field was initially three times greater than to matched T. parviflora plants, apparently as a consequence of adult oviposition preference. Subsequent larval migration from defoliated to green plants resulted in roughly equivalent defoliation of both hosts. Where the two Tamarix species co-occurred in northern Nevada, the ‘preference’ for T. ramossisima was apparent because it was consistently more heavily colonized, and utilization of T. parviflora declined as insect densities diminished. These results, and the lack of alternative explanations for establishment failures at T. parviflora sites (predation, developmental constraints, climate conditions), lead to the conclusion that such failures are based on host specificity as a consequence of adult behavioral avoidance of T. parviflora. They also suggest that risks to non-target plants both within the genus Tamarix (e.g., T. aphylla (Linneaus) Karsten, 1882) and in a related family (Frankeniaceae: Frankenia spp.) are low and should not delay program implementation.