SYNOPSIS In this paper an attempt is made to ascertain with what degree of accuracy records of the total radiation, or heat energy, received on a horizontal surface directly from the sun and diffusely from the sky, may be used to determine the intensity of daylight illumination on a horizontal surface. The utility of this investigation is obvious, since continuous records of the total radiation received on a horizontal surface are now obtained by the Weather Bureau at some of its more important stations, and electric-lighting companies are employing methods involving heat energy measurements in determinations of the variability of daylight. From Abbot's normal solar energy curve, and atmospheric transmission coefficients for different wave lengths of light also due to Abbot, the ordinates have been computed for solar energy curves in atmospheres of different degrees of transparency, and with the sun at different zenith distances. The ordinates of energy curves for a Planckian distribution at temperatures corresponding to color-temperatures of skylight measured by Priest and others have also been computed, and combined with the ordinates of the solar energy curve to determine the energy distribution in the total radiation received on a horizontal surface. The results indicate that midday radiation is richer in luminous rags than the radiation that is received when the sun is near the horizon. Comparisons between photometric measurements of daylight and pyrheliometric measurements of the total radiation lead to the same result. They indicate, however, that if the radiation intensity on a horizontal surface, expressed in gram-calories per minute per cm.,2 is multiplied by 6,700, the result will give the illumination intensity on a horizontal surface in foot-candles within ± 5 per cent, giving values which near noon are too low and which are too high when the sun is near the horizon.