This report describes laboratory measurements of the coefficient of thermal expansion of ice in a direction perpendicular to the optic axis. The practical problem which prompted the study is interesting. If the ice pressure on a dam is calculated on the assumption that the ice surface remains plane, and that the pressure arises because thermal expansion is restrained, the pressures calculated are much too high. It seems that buckling of the ice sheet plays an essential part. The author proposes to use his expansion measurements in a theory, as yet undeveloped, which will include buckling. His figures agree, in general, with those of other workers; but in tests on artificial ice he finds a variation of 30 per cent in the coefficient, which he attributes to the presence of impurities and to the influence on them of the thermal history of the specimen.