Methods of illumination to simulate the daily sunlight irradiation pattern were studied in relation to photo-hydrogen production using the photosynthetic bacterium Rhodobacter sphaeroides RV. Three illumination patterns were compared, in which the light intensity was changed in 1, 3, or 6 steps. As a control, outdoor experiments were also carried out over a 3-d period in Tsukuba, August 1996. Outdoors, hydrogen production by Rba. sphaeroides RV was dependent on the sunlight intensity: the total volume of hydrogen produced per day varied from 14 to 28 l.m(-2), while the total light energy ranged from 5.5 to 6.4 kWh.m(-2).d(-1), The maximum hydrogen production rate was 2.8 l.m(-2).h(-1) under a 4.5-cm light path and the average light energy conversion efficiency was 1.1%. Indoors, the hydrogen production rate was found to be independent of the mode of illumination among the three patterns employed. The maximum hydrogen production rate was 3.3 l. m(-2).h(-1) with a light energy conversion efficiency of 1.0%, and it was concluded that the single-step illumination method provides an appropriate simulation of sunlight. Saturation of hydrogen production occurs during high light intensity around noon and this plays a key role in the simulation.