A fundamental strategy evolved by organisms in cold habitats, obliged to cope with ice and freeze injuries, is ice-binding proteins (IBP). These proteins directly interact with ice surfaces and serve to depress the freezing point of the body fluids, to inhibit ice recrystallization, or to promote ice nucleation. The delicate control over ice growth make these proteins applicable in any field that requires control over ice growth, namely, in cryopreservation of cells, tissues and organs, in cryosurgery and in agriculture, in the frozen food industry and in complex material engineering. A variety of IBPs have been studied in the past decade, differing in sequences, structures, and specific activities. Most IBPs are grossly classified to moderate and hyperactive according to their thermal hysteresis activity, which is the gap between the equilibrium melting point and the freezing point of an ice crystal grown in IBP solution. We have studied the ability of IBPs to shape ice crystals in distinguishable forms characteristic of the particular protein type. IBPs with moderate thermal hysteresis activities induce elongated bipyramidal crystals – often with well-defined facets – while hyperactive IBPs produce more varied crystal shapes, such as the “lemon-like” crystals typically observed with Tenebrio molitor IBP. These unique morphologies are frequently considered to be growth shapes. We conducted a systematic study of ice shaping in solutions containing a wide range of IBPs. We found that although ice crystals in solutions of moderate IBPs do indeed grow into their faceted shapes, in the presence of most hyperactive IBPs, ice melts into its final shape. These melting shapes result from the affinity of the hyperactive IBPs for the basal planes of ice. A 3D simulation of a melting process using a velocity profile that reflects melting perpendicular to the basal plane confirmed our findings. Our results show a clear difference in the ice shaping mechanisms of moderate and hyperactive IBPs. This study implies that basal plane affinity may be predicted by simple observation of an ice crystal. Understanding the differences between the interactions of the various IBPs and ice is important for the future utilization of these proteins in industry and medicine. Source of funding: M.B.D was supported by The Lady Davis Fellowship thrust and The Valazzi-Pikovsky Fellowship Fund. The research was funded by the National Science Foundation (NSF), the Israel Science Foundation (ISF), the European Research Council (ERC). Conflict of interest: None declared. maya.bar1@mail.huji.ac.il