Subdural electrodes made of stainless steel, which were believed to be unsuitable for recording slow potentials, can still record Bereitschaftspotential (BP) (Neshige, R., Lüders, H. and Shibasaki, H. Recording of movement-related potentials from scalp and cortex in man. Brain, 1988, 11: 719-736) and ictal DC shifts (Ikeda, A., Terada, K., Mikuni, N., Burgess, R.C., Comair, Y., Taki, W., Hamano, T., Kimura, J., Lüders, H.O. and Shibasaki, H. Subdural recording of ictal DC shifts in neocortical seizures in humans. Epilepsia, 1996b, 37: 662-674) sufficiently. In this study, therefore, the effects of different kinds of metals on slow potential recordings were reevaluated. First, slow electro-oculograms (EOGs) were recorded with 3 different levels of input impedance (200 M omega, 470 k omega and 10 k omega) of a DC amplifier by using surface electrodes made of silver (Ag), silver/silver chloride (Ag/AgCl) and stainless steel. Secondly, BP was recorded by using the above electrodes with a long time constant of 3 s and with a fixed input impedance of 100 M omega. As a result: (1) slow EOGs were equally recorded with the input impedance of 200 M omega and 470 k omega regardless of the kind of metals used, although stainless steel electrodes caused baseline fluctuation, (2) low input impedance of 10 k omega allowed only the Ag/AgCl electrode to record slow EOGs without any decay, and (3) electrodes made of stainless steel could record BP as efficiently as the other two types of electrode with high input impedance. In conclusion, electrodes with a large surface area contact such as cup electrodes and an amplifier with a large input impedance, electrodes made of Ag, and even of stainless steel, can record slow potentials reasonably well.