Previously we have determined, with a full field of view, the relative contributions of scotopic and photopic luminance to pupil size at light levels typical of building interiors. Those studies were carried out in a white room with uniform reflectance, and with the viewed surfaces having an approximately uniform luminance distribution. To enhance the usefulness of the past results to lighting practice, we have constructed a simulated office where the viewed walls can have one of four very different colors, with quite different luminance distributions. This allows examination of interaction effects between wall spectral reflectivity and light source spectral distribution.

In the present study pupil sizes were obtained while subjects were viewing a very small screen television. A remote pupillometer was used, allowing subjects to sit in a comfortable chair without the inconvenience of chin rests or head gear. Seventeen subjects between the ages of 27 and 47 years were studied using illumination provided by conventional lamps, either a WW or a daylight fluorescent lamp. Pupil size variation was predicted by the value of the scotopic vertical illuminance at the eye. Even though the WW lamps are 50% more efficacious than daylight lamps in terms of photopic lumens per watt, daylight lamps can be as much as twice as efficacious in eliciting pupil size.