The transfer of plant-produced insecticidal Cry-proteins in the arthropod food web can be affected by a number of environmental and ecological factors. Despite this fact, most studies documenting multitrophic Cry-protein acquisition patterns in arthropods are conducted under controlled conditions whereas the number of field studies is limited. Such field studies, however, are valuable to understand multitrophic allocation dynamics of Cry-proteins under ecologically realistic conditions and are therefore important for the interpretation and design of laboratory hazard studies on genetically engineered (GE) crops. We thus sampled arthropods and plant structures in a field planted with GE dual-gene cotton plants producing the Cry-proteins Cry1Ac and Cry2Ab from Bacillus thuringiensis Berliner over the growing season. Cry-protein concentrations in field-collected plants, herbivores, and predators were quantified and compared with arthropods subjected to tri-trophic laboratory feeding assays. Both, field studies and laboratory assays showed that Cry-protein concentrations strongly decreased with increasing trophic level to values mostly below the detection limit in predators. Under field conditions, in-planta Cry-protein concentrations varied between plant structures and over the season. Concentrations in arthropods were mainly associated with feeding mode, feeding location on the plant, and to a lesser degree seasonality. Compared to plants, arthropods showed lower Cry2Ab:Cry1Ac ratios indicating that Cry2Ab might be less stable than Cry1Ac. Of all predators collected in the field study, we measured highest Cry-protein levels in jumping and crab spiders, predatory flies and some predatory hemipterans. This emphasizes the relevance of these groups for the risk assessment of GE cotton.