the species with different maize materials containing insecticidal Cry proteins from Bacillus thuringiensis (Bt). The results of those studies were often difficult to interpret, because only one GE plant was compared to one related non-GE control. In such a setting, effects of the Cry proteins cannot be distinguished from plant background effects, in particular when the test species is nutritionally stressed. In the present study, we tested the suitability of three different maize materials, i.e., flour, leaves and pollen, from five diverse non-GE maize lines (including EXP 258, a breeding line that is closely related to a SmartStax Bt maize) as exclusive food sources for D. magna. The parameters recorded included survival, sublethal endpoints such as body size, number of moltings to first offspring, time to first offspring, number of individuals in first clutch, total number of clutches, total number of offspring, average number of offspring per clutch, and population measures such as net reproductive rate R0, generation time T and intrinsic rate of increase rm. The results showed that D. magna can survive, grow and reproduce when fed only maize materials, although the performance was poorer than when fed algae, which indicates nutritional stress. Large differences in life table and population parameters of D. magna were observed among the different maize lines. Our results suggest that confounding effects caused by nutritional stress and plant background might explain some of the conflicting results previously published on the effects of Bt crops on D. magna. Using 95% confidence intervals for the means of the five maize lines for all measured parameters of D. magna performance in our study, we captured the natural range of variation. This information is useful for the interpretation of observed differences in D. magna performance between a GE plant and its non-GE comparator as it helps judging whether observed effects are of biological relevance. If differences between a GE and comparator line are observed and their biological relevance needs to be assessed in future risk assessments of GE maize, 1) the data on natural variation of the different parameters generated by previous studies can be informative (e.g. data from our study for maize fed D. magna); 2) for additional experiments the inclusion of multiple unrelated non-GE comparators should be considered; In addition, it should be taken into account that nutritional stress can affect the outcome of the study.
