The primary object of the work was the re-determination, by an electrical method, of the constant of Bunsen's ice-calorimeter. The heat was supplied by a manganin coil wound on a mica rack which fitted the interior tube of the calorimeter, and the results are based on determinations of E.M.F. and resistance. The current was adjusted so that the difference of potential at the ends of the heating coil was exactly equal to the E.M.F. of a number of standard cadmium cells n series. The conditions were varied as much as possible. Thus the rate of energy supply in the fastest experiments was more than seven times that in the slowest. Errors due to progressive freezing or thawing of the ice mantle were greatly diminished by suspending the calorimeter within a transparent vacuum vessel of cylindrical form, the stem and capillary of the calorimeter projecting through a rubber stopper, and the vacuum vessel being completely embedded in powdered ice. The mean value of the calorimeter constant was found to be 15.486 milligrammes of mercury per mean calorie. The Constancy of the Density of Ice. - Various observers have advanced evidence tending to show that the density of ice at 0°C. is not a definite constant. A consideration of their work leads to the conclusion that the small variations of density found for different samples might be simply due to the presence of occluded water or an amorphous modification cementing the ice crystals together. The value (80.30) of the latent heat of fusion of ice, calculated from the ice calorimeter, supports this view, as it is higher by about 0.7 per cent. than the value obtained by direct determinations with ice in bulk.