The gas-ion-feedback electron multiplier has been recognized for some time as capable of yielding a large-area electron source whose output can be spatially controlled on the scale of a picture element. Such a high-gain multiplier array can be conveniently formed using foils of oxidized Al-Mg alloy bonded to glass plates. During operation of the display the MgO dynodes of the electron multiplier are subject to intense ion and electron bombardment. This affects the device through three kinds of interaction. The ultimate life of a dynode is limited by sputtering. However, for the ion doses encountered during 104h of operation, the changes in secondary emission are not related primarily to sputtering damage but rather to the physisorption of the gas atoms on the MgO surface. This entrapment of gas also causes changes in the ambient pressure which must be compensated for. Finally, electron bombardment by itself can cause dissociation of MgO unless the MgO film has been stabilized by making the surface stoichiometric. Stabilized MgO dynodes can withstand a primary electron dose of 5000 C/cm2without damage and a dose of A+or He+ions of about 2 × 1017ions/cm2for an allowable decrease in secondary emission of 30 percent. An analysis of multiplier operation shows that a multiplier life of at least 104h can be expected in argon and helium; helium is the preferred choice because of the higher operating pressure and smaller sputtering damage to the electrodes. The sticking coefficient of 500-eV He ions on clean MgO at 23°C was measured to be 0.4 and it was found that at least six monolayers of He could be pumped into the dynodes. The characteristic time for gas release from the walls τ r is not constant but depends on the ion energy and on the degree of inward diffusion during the experiment. At 23°C, \tau_{r} > 100 h initially for 500-eV ions, and only 25 percent of the gas is recovered after three days. At 90° C, \tau_{r} \approx 1 h. These data were used to compute the pressure variations during various operating cycles. It is concluded that a gas supply is needed which is capable of delivering a total quantity of 20 1. torr of He on demand over hundreds of hours at 10-3torr, and which can reabsorb a comparable quantity of gas in about a minute. A pumping speed as small as 0.2 1/s would be sufficient to accomplish this.