When laboratory host specificity tests on weed biological control agents produce ambiguous results or are suspected of producing false-positive findings, field cage or open field tests can be used to help determine the true ecological host range of the agent. The leaf beetle Diorhabda elongata (Brullé) from Crete, imported to the United States for the control of saltcedar (Tamarix spp., Tamaricaceae), showed a low but variable ovipositional response to nontarget Frankenia spp. (Frankeniaceae) in previous laboratory tests conducted in small cages, where up to 11.4% of eggs were laid on these native plants. Results from field tests presented in this article show that no eggs were laid on Frankenia palmeri S. Watson and significantly more eggs were always laid on Tamarix ramosissima Ledebour than Frankenia salina (Molina) I. M. Johnston. Furthermore, the ovipositional response to F. salina was substantially lower than that recorded in laboratory tests. The percent of eggs laid on F. salina in field tests was 3.7 in a paired choice cage test, 4.3 in a multiple choice cage test, and 2.5 in a multiple choice open field test, suggesting that the true acceptance rate of the nontarget by D. elongata in the field will be lower than laboratory tests predicted. However, some damage was caused to F. salina by adult and larval feeding in the field, although this occurred only at the very end of the open field test, when D. elongata densities were extremely high, and all of the surrounding saltcedar had been totally defoliated. Scientific representatives from various stakeholder organizations (state, county, university, and environmental groups) viewed the open field test when in progress and reviewed the final results before advising State regulatory agencies on beetle redistribution. These test results, and the open review process, led regulators to conclude that redistribution of D. elongata in California was warranted owing to its significant ability to defoliate saltcedar, and its low rate of feeding on nontarget Frankenia spp. The introduction of D. elongata provides an interesting case study for risk assessment of a potentially efficacious weed biocontrol agent that may also be capable of using nontarget native plants.