Matthews et al. (1986) hypothesised that rock platforms around a short-lived ice-dammed lake margin in Jotunheimen, southem Norway, bad been rapidly eroded mainly through frost weathering associated with lake-ice development. They proposed a general model accounting for the development of the rock platforms in terms of deep penetration of the annual freeze-thaw cycle, the movement of unfrozen lake water towards the freezing plane, and the growth of segregation ice in bedrock fissures below lake level. This paper presents a test of this hypothesis by observations of the shoreline of the present-day lake, which has been maintained at a lower, stable leve! since about A.D. 1826 when the ice dam was removed. The presence of cliff and platform development at the present lake shore supports and improves the hypothesis. For the modem platform, width measurements (mean 3.6 m, range 1.5-5.75 m) are similar to those for the relict platform, whereas calculated erosion rates (mean 2.2 cm/year, range 0.9-3.6 cm/year) are overall slightly lower. The depth of water (0.9 m) at the cliff-platform junction suggested for the formation of the relict platform is modified to 0.6 m in the light of the present results. Implications of the results for the formation of the Norwegian strandftat are discussed.