ORIGINAL ABSTRACT

Glyphosate (GLY) is an herbicide widely used in agriculture. First considered as non-toxic or slightly toxic to bees, GLY and its different formulations have shown, more recently, to affect negatively the survival, development and behavior of these insects, even when used in doses and concentrations recommended by the manufacturer. Thus, the results of research on the toxicity of GLY to bees are often conflicting, which makes a meta-analysis interesting for data integration, generating a statistically reliable result. Therefore, this study aimed to evaluate the GLY effects on mortality of bees through a meta-analysis. For this, a search was carried out in the databases Web of Science, CAPES (Coordination for the Improvement of Higher Education Personnel - Brazil), Scopus, and PubMed. Papers that evaluated the effect of GLY on bee mortality published between 1945 and October 2020, were considered. After obtaining the data, R software was used to perform the meta-analytical tests. Sixteen papers on mortality were selected with 34 data sets. Most of the sets demonstrated differences between the control and experimental groups, showing that the treatments with GLY caused higher mortality of bees. The results considering the methodology used (ingestion or contact), the phase of the biological cycle (adults or larvae), and the dose (ecologically relevant dose and recommended by the manufacturer) were different when compared with their respective control groups. Therefore, GLY can be considered toxic to bees. It is important to emphasize that this meta-analysis identified that papers assessing the toxicity of GLY to bees are still scarce, for both lethal and sublethal effects, mainly for stingless and solitary bee species.