Résumé : Use of equipment-free, "manual"blood cultures is still widespread in low-resource settings, as requirements for implementation of automated systems are often not met. Quality of manual blood culture bottles currently on the market, however, is usually unknown. An acceptable quality in terms of yield and speed of growth can be ensured by evaluating the bottles using simulated blood cultures. In these experiments, bottles from different systems are inoculated in parallel with blood and a known quantity of bacteria. Based on literature review and personal experiences, we propose a short and practical protocol for an efficient evaluation of manual blood culture bottles, aimed at research or reference laboratories in low-resource settings. Recommendations include: (1) practical equivalence of horse blood and human blood; (2) a diverse selection of 10 to 20 micro-organisms to be tested (both slow- and fast-growing reference organisms); (3) evaluation of both adult and pediatric bottle formulations and blood volumes; (4) a minimum sample size of 120 bottles per bottle type; (5) a formal assessment of usability. Different testing scenarios for increasing levels of reliability are provided, along with practical tools such as worksheets and surveys that can be used by laboratories wishing to evaluate manual blood culture bottles.