Experiments showed that, as an ice specimen evaporated by exposure to a stream of dry nitrogen, a temperature gradient was created in the surface of the specimen, which became electrically charged. The charge was positive if the surface of the specimen was warmer than its interior, and negative if it was colder. A typical result was that a specimen of surface area 37 cm2 and internal temperature −20C exposed to a nitrogen stream of temperature −30C and velocity 15 cm sec−1 acquired a temperature gradient of magnitude 90C cm−1 in its surface and became negatively charged at a rate i=3.3 × l0−4 esu of current. These observations are explicable qualitatively in terms of the Latham-Mason theory of charge transfer associated with temperature gradients in ice. Calculations indicate that it cannot be determined whether quantitative agreement exists until measurements have been made of the distribution of temperature and charge within the volume of ice specimens along which temperature gradients exist. A... Abstract Experiments showed that, as an ice specimen evaporated by exposure to a stream of dry nitrogen, a temperature gradient was created in the surface of the specimen, which became electrically charged. The charge was positive if the surface of the specimen was warmer than its interior, and negative if it was colder. A typical result was that a specimen of surface area 37 cm2 and internal temperature −20C exposed to a nitrogen stream of temperature −30C and velocity 15 cm sec−1 acquired a temperature gradient of magnitude 90C cm−1 in its surface and became negatively charged at a rate i=3.3 × l0−4 esu of current. These observations are explicable qualitatively in terms of the Latham-Mason theory of charge transfer associated with temperature gradients in ice. Calculations indicate that it cannot be determined whether quantitative agreement exists until measurements have been made of the distribution of temperature and charge within the volume of ice specimens along which temperature gradients exist. A...