The operating problems of supercalenders and previous studies of supercalendering mechanisms are reviewed. In spite of observations which suggest random variations in heat source strength a one-dimensional periodic heat source model which describes the heating-up of a supercalender filled roll is found to approximate radial temperature distributions, measured at a number of combinations of speed and load, within filled rolls. Subsurface de format ions in a transaxial plane of a filled roll and the rolling friction of supercalender nips have been measured over a range of speed and load. Friction torque varies as √P³ where P is the nip load per unit axial length of roll contact. Ho correlation between friction and speed or roll temperature was revealed. Tnese measurements, as well as those of filled roll surface topography and hardness, indicate that friction and roll heating are surface phenomena dependent on random deformations of the filled roll.