BACKGROUND: Controlled blood-feeding is essential for maintaining laboratory colonies of disease-transmitting mosquitoes and investigating pathogen transmission. We evaluated a low-cost artificial feeding (AF) method, as an alternative to direct human feeding (DHF), commonly used in mosquito laboratories. METHODS: We applied thinly-stretched pieces of polytetrafluoroethylene (PTFE) membranes cut from locally available seal tape (i.e. plumbers tape, commonly used for sealing pipe threads in gasworks or waterworks). Approximately 4 ml of bovine blood was placed on the bottom surfaces of inverted Styrofoam cups and then the PTFE membranes were thinly stretched over the surfaces. The cups were filled with boiled water to keep the blood warm (~37 degrees C), and held over netting cages containing 3-4 day-old inseminated adults of female Aedes aegypti, Anopheles gambiae (s.s.) or Anopheles arabiensis. Blood-feeding success, fecundity and survival of mosquitoes maintained by this system were compared against DHF. RESULTS: Aedes aegypti achieved 100% feeding success on both AF and DHF, and also similar fecundity rates (13.1 +/- 1.7 and 12.8 +/- 1.0 eggs/mosquito respectively; P > 0.05). An. arabiensis had slightly lower feeding success on AF (85.83 +/- 16.28%) than DHF (98.83 +/- 2.29%) though these were not statistically different (P > 0.05), and also comparable fecundity between AF (8.82 +/- 7.02) and DHF (8.02 +/- 5.81). Similarly, for An. gambiae (s.s.), we observed a marginal difference in feeding success between AF (86.00 +/- 10.86%) and DHF (98.92 +/- 2.65%), but similar fecundity by either method. Compared to DHF, mosquitoes fed using AF survived a similar number of days [Hazard Ratios (HR) for Ae. aegypti = 0.99 (0.75-1.34), P > 0.05; An. arabiensis = 0.96 (0.75-1.22), P > 0.05; and An. gambiae (s.s.) = 1.03 (0.79-1.35), P > 0.05]. CONCLUSIONS: Mosquitoes fed via this simple AF method had similar feeding success, fecundity and longevity. The method could potentially be used for laboratory colonization of mosquitoes, where DHF is unfeasible. If improved (e.g. minimizing temperature fluctuations), the approach could possibly also support studies where vectors are artificially infected with blood-borne pathogens.