Search

A main point of our recent paper (Lovei et al. 2009) is that there are non-neutral effects of Cry toxins and proteinase inhibitors (PIs) on natural enemies in the laboratory and that the pattern of responses is complex and needs additional analysis. Shelton et al. (2009) aggressively attacked this conclusion. They claimed that all negative effects of Cry toxins are caused by effects of sublethally affected hosts and prey. We suggested in Lovei et al. (2009) and reiterate here that the actual situation is not that simple when laboratory studies are considered. We made our point by using statistical meta-analysis to show that there are more nonzero effects of Cry toxins and PIs on natural enemies than expected under a statistical null hypothesis that all observed effects were zero. The interested reader may want to examine the longer history of some of these issues (Lovei and Arpaia 2005; Andow et al. 2006; Romeis et al. 2006a,b). In our rebuttal, we first address the deeper, fundamental questions raised by Shelton et al. (2009) about the value of meta-analysis and then proceed to rebut the core criticisms about our statistical methods. Although Shelton et al. (2009) raised many other issues, we limited our rebuttal to these central issues; our lack of comment does not imply agreement with their other complaints. Shelton et al. (2009) make two criticisms of our work that are, in actuality, more fundamental criticisms of meta-analysis. These criticisms are made, in part, to defend the methods used and conclusions reached in reviews by O'Callaghan et al. (2005) and Romeis et al. (2006b), neither of which are based on meta-analyses. First they argued that nonsignificant P values are "devoid of futher meaning and interpretation" (Shelton et al. 2009, p. 318), and second, they …

Tri-trophic impacts of transgenicBacillusthuringiensis (Bt) cotton GK12 and NuCOTN 99B were studied using a predator, the great lacewing Chrysopa pallens (Rambur), and its prey, the cotton aphid Aphis gossypii Glover, in laboratory feeding experiments. The parental nontransgenic cotton cultivar of GK12 was used as control. The predator was fed with uniform (aphids from a single cultivar) or mixed prey (aphids from the three cotton cultivars provided on alternate days). Mortality and development of the immature stages, pupal body mass, adult sex ratio, fecundity, and egg viability ofC. pallens were measured. When fed GK12-originated aphid prey, pupal body mass ofC. pallens was signiÞcantly higher than that of the control, more females emerged, and these females laid signiÞcantly more eggs. Other parameters were not impacted. Females emerging from larvae maintained on NuCOTN 99B-originated prey laid fewer eggs than those maintained on GK12. Other measurements did not differ signiÞcantly between the two Bt cotton cultivars. Compared with the control, mixed feeding signiÞcantly prolonged pupal development time and increased pupal body mass and per- centage of females but did not affect other parameters. These results indicate thatC.pallens is sensitive to aphid prey from different cotton cultivars. Transgenic Bt cotton GK12-originated aphid prey has no adverse impact on survival, development, and fecundity of C. pallens. Between the two Bt cotton cultivars, NuCOTN 99B-originated aphid prey provided to C. pallens in the larval stage may lower female fecundity. Mixed feeding of C. pallens with the two Bt cottonÐ originated prey and non-Bt prey may have some adverse impacts on pupal development.

Previously, we demonstrated via meta-analysis that Cry toxins and proteinase inhibitors that are used in GM plants have nonzero direct and indirect effects on arthropod predators and parasitoids in the laboratory (Andow et al. 2009, Lovei et al. 2009). Our purpose was not to analyze the risks associated with any particular GM plant, but to measure whether there are detectable nonzero effects and the implications for assessing risk. These conclusions were criticized by Shelton et al. (2009a, b) on statistical grounds. We have argued that mean effect sizes cannot be used to claim the lack of an effect of Cry toxins on diverse species as was done in previous meta-analyses (cit. Andow et al. 2009). Instead, one must measure whether the distribution of observed effect sizes are normally distributed, as predicted under the null hypothesis that none of the observed effects are significantly nonzero. We provide here a technical clarification of our calculations, respond to two comments by Shelton et al. (2009b), and clarify the theory underlying our statistical tests. ### Statistical Issues: Theory and Methodological Details. There remain some misunderstandings about the structure of our statistical model for meta-analysis and the nature of the null hypothesis. Meta-analysis can be conducted in two fundamentally different cases. In Case I, which is most common in medical meta-analysis, data are aggregated from several studies measuring the same response on the same species by using the same laboratory protocols. Meta-analysis theory assumes that the real response underlying the experimental observations is either positive, negative, or zero, but by necessity, it has a single value. Under Case I, the null hypothesis is that if there is one real response, μ, then the normalized data have a standard normal distribution with a mean equal to the single real response value, $$ \text H_0 :g \sim \text N(\mu ,1),\,\text {for …

This review uses a data-driven, quantitative method to summarize the published, peer-reviewed literature about the impact of genetically modified (GM) plants on arthropod natural enemies in laboratory experiments. The method is similar to meta-analysis, and, in contrast to a simple author-vote counting method used by several earlier reviews, gives an objective, data-driven summary of existing knowledge about these effects. Significantly more non-neutral responses were observed than expected at random in 75% of the comparisons of natural enemy groups and response classes. These observations indicate that Cry toxins and proteinase inhibitors often have non-neutral effects on natural enemies. This synthesis identifies a continued bias toward studies on a few predator species, especially the green lacewing, Chrysoperla carnea Stephens, which may be more sensitive to GM insecticidal plants (16.8% of the quantified parameter responses were significantly negative) than predators in general (10.9% significantly negative effects without C. carnea). Parasitoids were more susceptible than predators to the effects of both Cry toxins and proteinase inhibitors, with fewer positive effects (18.0%, significant and nonsignificant positive effects combined) than negative ones (66.1%, significant and nonsignificant negative effects combined). GM plants can have a positive effect on natural enemies (4.8% of responses were significantly positive), although significant negative (21.2%) effects were more common. Although there are data on 48 natural enemy species, the database is still far from adequate to predict the effect of a Bt toxin or proteinase inhibitor on natural enemies.

Larvae, pupae, and adults of ladybird beetles were observed in an insecticide-free and an insecticide-treated block in an apple orchard near Budapest, Hungary during 1980–1982. Species numbers and densities in all developmental stages were higher and species turnover was lower in the insecticide-free block. Diversity and similarity indices showed more diverse and constant assemblages in the insecticide-free block. The adult assemblage was the most rich in species, with 19 species in the insecticide-free and 12 species in the treated block. The most common species overall were Coccinella septempunctata, Adonia variegata, Adalia bipunctata , and Exochomus quadripustulatus . The larval assemblage included nine and eight species for the untreated and treated blocks respectively, whereas the pupal assemblage had eight species in both blocks. Immature stages of 11 species were found in at least 1 yr and 10 of these was able to complete development in the orchard. The rate of success was higher in the insecticide-free block for most species. Only A. bipunctata reproduced successfully in both blocks in all three years but A. decempunctata and C. septempunctata were often successful.