We present evidence that the production of secondary ice particles during rime growth on a moving modgy is greater at low velocities relative to the number of drops accreted (over the range 1.4 to 3 m s-1). This favours the hypothesis that 'splintering' is due to the formation, round the periphery of the drop, of an ice shell which later shatters. Shell formation is more likely if drops land lightly and do not spread over the ice substrate. |