SUMMARYA laboratory investigation of electric charge transfer during the impact of vapour-grown ice crystals and supercooled water droplets upon a simulated soft-hailstone target has shown that the magnitude of the charge transferred to the riming surface when crystals separate from it is a function of temperature, crystal dimension, relative velocity, liquid water content, and impurity content of the water droplets and hence the impurity content of the riming target. The sign of the charge transfer depends on temperature, liquid water content and droplet and rime impurity content.In the absence of crystals, no charge transfer was detected during riming. In the absence of supercooled water droplets, crystals impacting at 10m s-on an evaporating rime target produced a small negative charge on the rime of less than -0.25fC per separating crystal. When the target surface grew by vapour diffusion it gained a small positive charge during such interactions. Much larger charges and completely different charge transfer behaviour was noted during riming. The target became positively charged at high liquid water contents and temperatures above a critical value, but negatively charged at lower temperatures or with lower liquid water contents. The critical sign reversal temperature at a liquid water content of 1 g m -3 was about -20°C. At -10°C with a liquid water content of 2gm-3, a 125pm crystal impacting at 3ms-I charged the target by + lOfC upon separation. The charge transfer increased sharply with impact speed and crystal size. Warming the positively charging rime to cause it to evaporate failed to reverse the sign of the charge transfer. Experiments with impurities showed that the sign reversal temperature increased if the droplets contained contaminants at concentrations found in cloud water.It is suggested that there are two distinct charge transfer processes during crystal interactions with an ice target, the dominant one requiring the presence of supercooled water droplets. Careful control and knowledge of the microphysical properties of the clouds used in these experimental simulations has permitted an examination of charge transfer under many of the conditions used in previous studies. The results provide an understanding of the differences and a reconciliation between some of the previously disparate findings in terms of the two distinct charge transfer regimes.