Detection of stored-wheat beetle species and estimation of population density using unbaited probe traps and grain trier samples

Field studies were conducted in order to compare the effectiveness of unbaited probe traps and grain trier samples in the detection of several beetle species. On June 15, 1997. fourteen probe traps were placed in three steel bins containing wheat in Central Greece. Two of them were inserted in the central zone of the bulk, five in the half of the bin radius (median zone) and seven near the bin wall (distal zone). From June 30 up to and including January 1998, the traps were checked for adult beetles every fifteen days (15 sampling dates in total). On each date, samples were also taken from spots adjacent to trap locations, using a grain trier. The mean temperatures recorded near the bulk surface decreased by an average of 1.16 degreesC per sampling date. Twenty-four species, belonging to 14 families of Coleoptera were found. The most abundant species in the traps were Cryptolestes ferrugineus and Tribolium castaneum, while Sitophilus oryzae was the most abundant in the samples. Significantly greater numbers of adults were found in traps than in samples, during the whole sampling period; traps also detected adults in most locations while in the corresponding samples no adults were found. Significantly higher numbers of adults were found in the central sampling zone, using both devices. Taylor's power law parameters showed that in both sampling methods the most abundant species showed aggregated spatial patterns. Although the correlation coefficient values between trap catches and number of adults per trier sample differ significantly from zero (P < 0.01) in a high number of sampling dates, it can be concluded that traps are poor indicators of population density. The accuracy provided as a percentage of the mean, decreases exponentially with the increase of the mean value. Based on traps, the desired accuracy level (35%) is not achievable even in higher mean values or even with high numbers of traps. On the contrary, with trier samples it is possible to estimate population density with a relatively low number of sampling units. Significant differences in the required number of sampling units (sample size) were noted among species.