We conducted this study to assess the impact of the bait station used in the Carambola fruit fly eradication program on non-target insects indigenous to French Guiana. A sampling device was developed based on the bait station, using four treatments: fiberboard blocks impregnated with methyl eugenol (ME) and malathion (MA), ME only, MA only, and a non-impregnated block. A total of only 149 non-target insects were captured during the 52 weeks of the experiment. The non-target taxa captured included representatives of nine orders and 26 families. Analysis based on a Poisson model of captures indicated no difference in the taxa or number of specimens captured among the treatments, providing evidence that the bait stations used by the regional program had no greater impact on non-target insect populations than the non-impregnated blocks. Our results indicate that captures of non-target insects could he related to accidental exposure to the sampling device and as such can be considered random events. The results of this experiment support the hypothesis that bait stations made with ME and MA can be used in an area-wide program without risk to non-target insect populations.