The author discusses the low temperature behavior of several samples of frozen soils taken from the dry valleys of Antarctica. The samples were composed of various clay minerals and had varying water contents. It is demonstrated that some of the water remains unfrozen and that there is a dependency between the unfrozen portion and the surface area of the sample. It was pointed out that problems arising from the unfrozen water content of soils are of great interest to CRREL researchers as is the analysis of ice cores from Greenland and Antarctica as a technique for establishing past climates and in predicting future climates.