HOW COME? Making ice -- out of hot water

Can warm water ever freeze faster than cold water? asks reader Mary Drayer. Please explain why hot water freezes faster than cold water, writes reader Kris Darragh.

Most people, if forced to make ice the old-fashioned way -- by filling and then freezing an ice-cube tray -- would start with the coldest water available. After all, water freezes because it loses heat. And cold water has already lost more heat than warmer water. So it's just that much closer to the freezing point of 32 degrees.

But as it turns out, given the right conditions, cold water may actually freeze more slowly than hot.

Aristotle noted the paradoxical hot water phenomenon back in 4th century BC. Later, scientist Francis Bacon, born in 1561, and mathematician René Descartes, born in 1596, also discussed the unusual effect. But by the 20th century, scientific discussion of the freezing effect had all but ended.

However, in 1963, a boy named Erasto Mpemba, making ice cream in school, noticed that mixtures made with hot milk seemed to freeze faster than those made with cold. Later, in high school in Tanzania, Mpemba asked his teachers for an explanation. A visiting professor of physics, intrigued, did a series of experiments in his lab, apparently confirming Mpemba's observations. And in 1969, scientist Denis Osborne and Mpemba jointly published their findings.

Is the "Mpemba effect" real? Some experiments seem to show that hot water will sometimes freeze faster than less-hot water, depending on the starting temperatures, containers and freezing conditions.

According to one argument on how the effect might work, given two open containers of hot and cool water left in a freezer, the hotter water will evaporate faster, leaving less water to cool down. Also, evaporation makes water lose heat quickly. So an open container of warmer water may freeze first. Meanwhile, a thin layer of ice will quickly form on the surface of the cooler water, insulating the water from further evaporation -- and actually slowing the rate of ice formation under the surface.

In 2010, scientist James Brownridge of Binghamton University published the results of his own experiments. He compared tap water at 212 degrees (water's boiling point) to distilled water at 77 or lower. Repeating the experiment 27 times, Brownridge found that the hot tap water always froze first.

He notes that since the two types of water have slightly different freezing points, he hasn't proved the Mpemba effect in all cases.

But he did show that given two common kinds of water, a boiling-hot sample can actually freeze faster than a cool sample.