Sea ice is a crucial component of the climate system, affecting albedo, heat and gas exchange, deep-water formation, atmospheric chemistry and ocean biology. However it remains diffcult to model recent trends in both polar regions, and the dataset is very short for assessing the role of signal against noise. Obtaining good reconstructions of past ice extent would allow more rigorous assessment of sea ice models, and would better constrain climate models for past times. It would also allow an assessment of the role of sea ice feedbacks under a variety of changed climates and forcings. A number of proxies for sea ice are in use, but it remains challenging to reconcile the different data and draw up quantitative reconstructions. In this session we welcome contributions about the development and use of proxies for sea ice, about the use of sea ice modelling for previous times, and of course comparisons and joint use of models and data. We also welcome more general contributions about the role of sea ice feedbacks that draw on both palaeo and instrumental or satellite data, and that consider such feedbacks in the past, present and future.