Celsius: A temperature scale reversed in time
The Celsius temperature scale began in the opposite direction from its familiar modern form. In 1742, Swedish astronomer Anders Celsius set boiling water at 0 and freezing water at 100. His proposal made the hotter fixed point carry the smaller number. The modern scale reverses those positions, but the original work involved more than assigning endpoints. Celsius investigated the effects of atmospheric pressure on water's behavior, finding that the melting point of ice was essentially unaffected while the boiling point changed. He proposed calibrating the boiling point at mean barometric pressure at mean sea level.
The reversal developed through several people. French physicist Jean-Pierre Christin adopted the direction with freezing at 0 and boiling at 100 in 1743. On 19 May that year, he published the design for a mercury thermometer made by Pierre Casati, called the Thermometer of Lyon. Whether Christin independently developed that arrangement or reversed Celsius's work remains a matter of attribution. In 1744, the year Celsius died, botanist Carl Linnaeus also reversed the temperature scale. His greenhouse instrument was made by Daniel Ekstrom, whose workshop occupied the basement of the Stockholm observatory. Other scientists and instrument makers also received credit for independent development.
Linnaeus left a concrete glimpse of the scale in use. His dissertation Hortus Upsaliensis, defended on 16 December 1745 by his student Samuel Naucler as respondent, described temperatures inside the orangery at the University of Uppsala Botanical Garden. Sunlight through the windows could bring the hot greenhouse section to 30 degrees. The gardener generally kept it from rising above 20 to 25 degrees and from falling below 15 degrees in winter. This is the first known Swedish document reporting temperatures in the modern forward direction. Its setting places the developing measurement system among living plants and daily greenhouse management.
For many years, the temperature scale was called centigrade. The name combines Latin words for a hundred and steps, and scientific communities used degrees centigrade through the nineteenth century. In some languages, however, centigrade also referred to an angular measurement. Centesimal degree created another ambiguity related to angular units. International measurement authorities formally adopted degree Celsius in 1948, honoring the astronomer while removing that confusion. The older word persisted informally in French and English. Australian weather reports used Celsius exclusively after 1 September 1972, while BBC Weather changed its wording from centigrade to Celsius in February 1985.
The unit name can express either a particular temperature or an interval between temperatures. That double use means the context must make the meaning clear: the symbol alone cannot distinguish a measured point from a difference. A Celsius degree has the same magnitude as a kelvin, so an interval can be expressed using either unit. The scale nevertheless is relative, rather than an absolute ratio scale. An object at 20 degrees Celsius does not have twice the energy it has at 10 degrees Celsius, and 0 degrees Celsius is not the lowest possible temperature. Absolute zero corresponds to -273.15 degrees Celsius.
The temperature scale also has specific conventions for its written form. The full unit name is degree Celsius, with degrees Celsius as its plural. Since the kelvin replaced degrees Kelvin in 1967, degree Celsius has been the only SI unit with an uppercase letter in its full name. International Bureau of Weights and Measures rules put the number before the unit and separate them with a space. Unicode includes a single Celsius character for compatibility with older encodings, but recommends the ordinary degree sign followed by the capital letter C for normal use. These details distinguish a unit's name from the symbols used to print it.
Definitions changed as measurement developed. Freezing and boiling water anchored the scale throughout the nineteenth century and the first half of the twentieth. In 1948, the definition shifted to the triple point of water. A later refinement specified the water's isotopic composition. In 2019, a definition based on the Boltzmann constant severed the kelvin's defining dependence on water's properties. Celsius remains defined through the kelvin. Each formal change preserved the earlier numerical temperatures within the accuracy available at the time, even while the physical basis of the definition changed from water reference points to a constant.
Water's melting and boiling points consequently became measured quantities instead of defining fixtures. A boiling point measured under one standard atmosphere using an earlier calibration was 99.9839 degrees Celsius, while a high-precision instrumentation standard gave a slightly different value. Such differences have little everyday significance because boiling responds strongly to atmospheric pressure. Even an altitude change of 28 centimeters changes the boiling point by one millikelvin. A measurement made above sea level can therefore vary because of pressure, instrument uncertainty, and trace impurities in the water, each affecting the observed result.