In the experiment, the cathode was held at some voltage $V_C$ and the anode was held at a voltage $V_A>V_C$, so that electrons would be emitted from the cathode and strike the anode. In a vacuum, the amount of kinetic energy an electron arriving at the anode had would be proportional to the difference in voltage $V_A-V_C$. But the experiment was conducted in a tube filled with mercury vapor, with which electrons might collide and thus lose energy. The experiment was designed to measure this loss, so they needed some way to distinguish between electrons with different amounts of kinetic energy. This is where the grid comes in. The grid was held at a voltage $V_G>V_A$, and so electrons passing through the grid (it was assumed only a small fraction collided with it) did one of two things:
- If they had enough kinetic energy to overcome the potential difference $V_G-V_A$, they would continue on to the anode, thus contributing to the current between the cathode and the anode.
- If they had too little kinetic energy, they would eventually fall to the grid and contribute to the current between the cathode and the grid.
Thus by comparing the two currents, Frank and Hertz were able to determine what fraction of electrons had less than a certain amount of kinetic energy. By calculating the expected amount of kinetic energy the electrons would have in a vacuum, they were able to determine what fraction of electrons had lost a certain amount of energy to collisions. Since almost all of this energy loss was due to inelastic collisions which raised the energy level of n electron in the Mercury atom, they were able to predict this using the Bohr model. The agreement of these predictions with the experiment provided support for the Bohr model.