A mechanism for electric charge separation during freezing of slightly impure water, discovered by Workman and Reynolds (1950). When a very dilute solution of certain salts freezes rapidly, a strong potential difference is established between the solid and liquid phases. For some salts, the ice attains negative charge, for others, positive. This mechanism was thought to play a role in the charging of thunderstorms. It is now known that it cannot account for the electrification of riming graupel by the shedding of charged surface water because the droplets freeze faster than the time taken for a substantial ice– water freezing potential to occur. This mechanism has been suggested as one possible mode of thunderstorm charge separation in those portions of a thunderstorm downdraft where snow pellets or hail particles sweep out supercooled water drops. Partial freezing and partial blow-off of a liquid film could lead to charge separation. The acknowledged predominance of dry growth of graupel in New Mexico thunderclouds led Reynolds (1953) to question the viability of this mechanism in the atmosphere. This should not be confused with the Reynolds effect.