The principle cause of frost heave is the forma on of segregated ice-ice lenses-in freezing soil columns. Despite much experimental and theore cal work, there remain many ques ons about the fundamental process by which this occurs. Frost-heave models fall into two main classes: capillary and frozen-fringe models. Which model is appropriate depends on whether there is a frozen fringe; these are diffi cult to observe but some experimental evidence does exist. Recent advances have revitalized the capillary model, such as the engulfment model and the concept of geometrical supercooling. Key experimental and theore cal challenges remain to be resolved.Frost heave refers to the upward displacement of the ground surface caused by the formation of ice lenses-discrete layers of ice that form in freezing soil. h is phenomenon is partly responsible for beautiful surface patterns that appear in very cold, permafrost areas ( Fig. 1). Similar patterned ground has been observed on ice-rich portions of Mars, and the observations have been used to draw interesting conclusions about previous climatic conditions (Gallagher et al., 2011). Frost heave also has many practical and industrial implications. Residents in cold countries are familiar with the annual appearance of frostheave-induced road damage at er winter cold spells, and the heaving forces are capable of damaging infrastructure such as pipelines, railways, and buildings. In the United States, more than two billion dollars is spent annually repairing frost-heave damage to roads alone (DiMillio, 1999). h us, worldwide, the cost is tremendous. h ere are still many unanswered questions about the underlying mechanisms controlling frost heave. Almost a century has passed since Stephen Taber demonstrated experimentally the basic features of frost heave in soils (Taber, 1916(Taber, , 1929(Taber, , 1930, and scientists are still actively working to understand his observations. Taber's key result was to show that frost heave is not, as commonly assumed, caused by the expansion of water on freezing; heWe critically review the two main classes of frost heave models-capillary and frozen-fringe models. We describe recent developments in the fi eld and highlight areas where new research is needed.