Tonight, as we all know, the punt, the Deutschmark, several different ilks of continental franc, and a clutch of other currencies will disappear forever, to be replaced by a ubiquitous denomination called the euro. But it has all happened before in meteorology, and the science has managed to survive the change unscathed.
With the development of the thermometer as a scientific instrument, anyone who constructed such a device seemed to throw in a new scale of temperature for good measure. As a result, between the years 1640 and 1780 at least 35 different scales of temperature were proposed and came into limited use. The Fonlex Scale had a freezing point of -33 degrees and a boiling point of 245; on the Paris Scale the corresponding points were 24.5 and 238, on the Cruquis 106 and 151, and on the Edinburgh Scale, 8 and 47. Two others were constructed in such a way as to give diminishing numbers for increasing temperatures: the Royal Society Scale had a freezing point of 79 and a boiling point of -170 degrees, and the De Lisle ranged from 150 down to zero.
As with currencies in recent times, however, rationalisation came to be seen as clearly necessary. By the middle of the 20th century most of the scales had fallen into disuse and been forgotten, and there only remained the Celsius or Centigrade scale used almost universally in scientific work, and Fahrenheit which was widely used in English speaking countries. Then forty years ago, the final step took place in meteorology: in the early 1960s, virtually everywhere except in the United States, meteorologists adopted Celsius.
Apart from the mathematical convenience of going decimal, and the euro-like advantages of being, as it were, an almost universal currency, many meteorologists felt that the Celsius degree was a more useful tool in weather matters, and that the Fahrenheit degree - roughly half the magnitude of a degree Celsius - had been rather too precise for use in forecasting. Temperature is a very variable element; on a sunny day it may vary over short distances and times by several degrees Fahrenheit, and the somewhat coarser Celsius degree seemed a more appropriate means by which to specify it.
Another advantage seen in the Celsius scale was that it makes a highlight of the freezing point, as the dividing-line between positive and negative values of the temperature. The freezing point is of practical importance and those interested are more likely to be alerted to potential hazards by "minus 2", than on hearing the rather unremarkable figure in Fahrenheit of "28 degrees".