These flakes have been screened from a highly-fritted boron glaze mixed using hard water and stored for a year. They formed as a film across the top of the settled surface and on the walls of the bucket. Frits are stoichiometric, they should not dissollve, but they do. Perhaps this is evidence that frit makers are unable to precisely control production parameters.
These 1 mm-sized crystals were found precipitated in a couple of gallons of glaze containing 85% Ferro Frit 3195. They are cubical, hard and insoluble. Why and how to do they form? Many frits are slightly partially soluble and the degree to which they are are related to the length of time the glaze is in storage, the temperature, the electrolytes and solubles in the water and interactions with other material particles present. The solute then interacts with other materials particles to form insoluble species that crystallize and precipitate out as you see here. These crystals can be a wide range of shapes and sizes and come from leaded and unleaded frits.
Crystals or crystalline particles will often form over time in ceramic glaze slurries that contain slightly soluble materials.