Given what Gene wrote, it looks to me like the use of the Celsius scale is quite similar to the use of the liter, in that both are defined in terms of an SI unit, are not themselves units, and are kept mostly for reasons of history and convenience than anything else.
Ezra Gene Mechtly wrote: > What is the difference between the historical definition > of the centigrade temperature scale and the modern Celsius scale? > > In 1742, Anders Celsius published a paper titled (after translation) > "Observations on Two Constant Degrees on One Thermometer" in the > Transactions of the Swedish Academy of Sciences. > > The *two constants* were the ice point and the boiling point of water, > separated by exactly 100 degrees on that "centigrade" scale. Celsius > chose the boiling point as the zero on his original scale. > > On the other hand, the modern Celsius scale has *only one* fixed point, > a zero at 273.15 K, and an increment (unit) of temperature difference > identical to 1 K, as Jim Frysinger has explained. > > Thus, the original "centigrade" scale was, in fact, *different* from the > modern Celsius scale. > > However, the definition of the original centigrade scale, and its > immediate successor with the zero reset at the ice point, were effectively > discarded by a series of resolutions by the CGPM as Joe has explained. > > Gene.
