Electrically charged species in flames can be studied by sampling a flame through a metallic nozzle into a mass spectrometer. Here, a voltage has been applied between the sampling nozzle, together with its mounting plate, and a metallic burner to measure current-voltage characteristics, i.e., plots of a sampled ion current las registered by the mass spectrometer) vs the applied voltage. These characteristics are different for a large, wide, multi-tube (Meker) burner and a small, single-tube (Bunsen) burner. For diagnostic purposes, net current-voltage characteristics were also measured for the total current. positive or negative, collected by the burner or the sampling system. The major problem is to characterize, even qualitatively, the distribution of electrical potential between the burner and sampling system for a flame of H-2 + O-2 + diluent, doped with Cs, at atmospheric pressure. It is confirmed that such flames must be treated as a weakly ionized, flowing, collision-dominated (continuum) plasma, in which diffusion of charged species can be ignored. With the sampling system at a negative potential with respect to the burner, the flame conforms to a sheath-convection model, so that convection to a sheath of positive ions covering the sampling nozzle dominates, while electrons are repelled. With the nozzle, etc. at a positive potential, the applied voltages employed are too low to stop the positive ions and thereby create an electron sheath, but the hulk flame plasma can sustain a substantial ohmic potential difference, and charge collection is dominated by mobility considerations. Of key importance for different burners is the burner-flame plasma impedance, which can greatly affect the distribution of potential between the burner and the nozzle. Elucidation of these phenomena provides a better understanding of the electrical aspects of exactly what is being sampled, when ions are extracted this way from a flame.