If a sample of water is repeatedly cooled into the supercooled region until it eventually nucleates and
freezes, the resulting data on nucleation temperatures can be plotted as a survival curve. We look here
at the slope of such curves and suggest some reasons for both a changing slope, either with time or with
more repeats, and also for the slope being related to multiple nucleation sites, either at the same time
or perhaps subatomic particles, such as neutrons, passing through the sample and causing nucleation,
rather than an impurity or defect on the container wall.
