There are many organisms throughout the natural world that survive cold or freezing temperatures in th~ir habitat. As part of a variety of cold tolerance meqhanisms; many of these organisms generate Antifreeze Proteins, or AFPs, to provide protection against the growth of ice crystals in their tissues and cells. AFPs provide this defence by binding to the surface of ice crystals to modify, or prevent, ice crystal growth and re-crystallisation. In this thesis, a variety of physical methods are described that characterise AFP binding to ice at the molecular level. The aim was to define which planes on an ice crystal an AFP b~ds to, the effect this binding has on the ice microstructure and the resulting effect on crystal growth in solution. These data could provide a link between differences in the molecular structure of the various AFPs and the different effects they have on ice. Such information is important in ' t developing an understanding of physic~hemicalmechanisms ofAFP action and, could be useful in exploiting AFPs in a variety 'of food, agriculture and biotechnology applications. The binding ofAFPs has been probed by the single ice crystal hemisphere technique (Knight et al 1991). This technique and its calibration are described, in addition to its use in the identification of binding sites for several AFPs. The ice crystal binding site was identified using crystallographic principles, to enable the interaction between the protein moiety and the relevant plane to be proposed. .A number of previous workers have established that AFP action can be enhanced by the presence of other molecules in biological fluids. This effect has been probed and the influence of small molecular solutes, dyes and other proteins was observed directly using the ice hemisphere technique. The effect ofAFP binding on the ice crystalline microstructure is also reported and the consequences that these foreign particles have on ice crystallisation is discussed. Fundamental studies involving a variety of techniques including electron microscopy, X-ray tomography and material properties, provide crucial evidence for the explanation ofAFP action. The kinetics of ice crystal growth in an AFP solution are calculated from time-dependent studies of crystal size and shape within the hysteresis gap. The relationship between binding site and crystal shape was also explored via this experiment' and several different AFP shapes were recorded. With comparison to protein binding information, a causal model was established. These data are crucial to the development of theories to explain AFP action in vitro and in vivo. Each of the proposed AFP binding mechanisms is reviewed with respect to the evidence presented in the thesis and conclusions regarding AFP action reached.