The development of multistep nucleation theory has spurred on experimentalists to find intermediate metastable states that are relevant to the solidification pathway of the molecule under interest. A great deal of studies focused on characterizing the so-called "precritical clusters" that may arise in the precipitation process. However, in macromolecular systems, the role that these clusters might play in the nucleation process and in the second stage of the precipitation process, i.e., growth, remains to a great extent unknown. Therefore, using biological macromolecules as a model system, we have studied the mesoscopic intermediate, the solid end state, and the relationship that exists between them. We present experimental evidence that these clusters are liquidlike and stable with respect to the parent liquid and metastable compared with the emerging crystalline phase. The presence of these clusters in the bulk liquid is associated with a nonclassical mechanism of crystal growth and can trigger a self-purifying cascade of impurity-poisoned crystal surfaces. These observations demonstrate that there exists a nontrivial connection between the growth of the macroscopic crystalline phase and the mesoscopic intermediate which should not be ignored. On the other hand, our experimental data also show that clusters existing in protein solutions can significantly increase the nucleation rate and therefore play a relevant role in the nucleation process.prenucleation clusters | phase transition | self-purification T he process of crystallization is generally considered to occur in two consecutive but very different stages: nucleation and growth. The first stage was already studied more than two centuries ago by Gibbs, who considered the nucleation of water droplets from a supersaturated vapor through the formation of globulae. He was the first to develop a thermodynamic formalism of nucleation by considering the generation of nuclei of a liquid phase as a density fluctuation of the parent phase (1, 2). Since then, Gibbs' nucleation theory has been extended to the nucleation of solid phases from solution and gaseous phases from which the current paradigm (based on the capillary approximation) for nucleation emerged, i.e., the classical nucleation theory ]. There is, however, an increasing body of evidence that shows that CNT can fail drastically when used in cases where the implicit and explicit assumptions of CNT are poorly justified (for a full dissection of the limitations, see refs. 9-11). An obvious situation where CNT will have limited applicability is in cases where the old and the new phases differ by at least two order parameters, e.g., density and structure (12).Recent theoretical, computational, and experimental efforts have demonstrated that densification and local increase in crystallinity need not occur simultaneously (13)(14)(15)(16)(17)(18)(19)(20). These results have inspired the development of a new approach that considers nucleation from solution as a multistep process attributing key roles to metastabl...