When pure water is cooled at $10 6 K=s, it forms an amorphous solid (glass) instead of the more familiar crystalline phase. The presence of solutes can reduce this required (or ''critical'') cooling rate by orders of magnitude. Here, we present critical cooling rates for a variety of solutes as a function of concentration and a theoretical framework for understanding these rates. For all solutes tested, the critical cooling rate is an exponential function of concentration. The exponential's characteristic concentration for each solute correlates with the solute's Stokes radius. A modification of critical droplet theory relates the characteristic concentration to the solute radius and the critical nucleation radius of ice in pure water. This simple theory of ice nucleation and glass formability in aqueous solutions has consequences for general glass-forming systems, and in cryobiology, cloud physics, and climate modeling. DOI: 10.1103/PhysRevLett.110.015703 PACS numbers: 64.60.QÀ, 05.40.Àa, 64.70.dg Ice nucleation and growth are of major interest in fields ranging from cryobiology to atmospheric physics. Ice is a key issue in cryopreservation of cells and tissues [1] and in cryocooling of samples for macromolecular crystallography [2,3], where solutes like salts, sugars, alcohols, and polyols can have dramatic effects on ice formation. In atmospheric physics, models of cloud formation are sensitive to the nature of the critical nucleus of ice crystals [4] with implications for climate models [5]. Supercooled water is an interesting system in its own right [6], and the formation of crystalline phases from supercooled solutions is an active area of study [7].Previous experiments have focused on properties such as the melting and glass transition temperatures, and models to explain the data have largely been phenomenological. Similar models have been applied to explain glass formability in a wide variety ofnonaqueous systems. Here, we report measurements of the minimum cooling rates [or ''critical cooling rates'' (CCRs)] required to prevent ice formation in aqueous solutions during cooling to $100 K or below. We show that a surprisingly simple statistical modification to classical thermodynamic nucleation theory provides an excellent account of these data.We studied eight different solutes (see Fig. 1), including a salt (sodium chloride), simple alcohols (methanol, ethanol), sugars (dextrose, trehalose), polyols (glycerol, ethylene glycol), and polyethylene glycol 200 (PEG 200). All are compact and highly soluble and can have large effects on critical cooling rates required for vitrification.The effects of these solutes on ice nucleation were evaluated by measuring the critical cooling rate above which no ice was observed. Below the critical rate, a sample turns opaque on cooling, indicating the formation of polycrystalline ice. As the rate increases, a transition to transparent samples is observed. This optical transition corresponds to a transition in the x-ray diffraction patterns obtained from the cooled sam...