To select efficient baits to attract fruit flies (Diptera: Tephritidae), the use of field cages represents a good compromise between laboratory and outdoor studies. Nevertheless, the methodological details of such experiments up to now have had little attention. To assess the influence of intrinsic and extrinsic factors on the efficiency of food attractants for Bactrocera cucurbitae (Coquillett), a series of methodological experiments were conducted in field cages with McPhail traps. Baited traps should initially be randomly placed and furthermore regularly rotated to lower the influence of climatic conditions. During an experiment lasting 8 h, the presence of already captured flies did not influence the attractiveness of the traps. The presence of potted host plants in the field cage allowed for better dispersion of the flies and enhanced the discrimination potential of the experiment. Moreover, yellow traps should be painted black to limit visual bias. Finally, in the conditions of these experiments, the sex of tested flies and the diet given to them during rearing, with or without protein, had no influence on the qualitative results of choice experiments. However, these factors greatly influenced the total amount of captured flies: protein-deprived females were more responsive than all others. Furthermore, whether the tested flies were sexually mature had a significant influence on their responsiveness to protein baits. These results are discussed to establish recommendations for further field cages experiments.