msb@sq.uucp (Mark Brader) (08/16/87)
Sci.misc has just seen a bunch of postings about the origins of the Fahrenheit temperature scale. What I remembered differed from some of them, and I was curious, so I decided to look it up. (!) In my public library I found the book: A History of the Thermometer and its Uses in Meteorology by W.E. Knowles Middleton the Johns Hopkins Press, Baltimore, 1966 And this is what I found out (slightly augmented from other sources): Daniel Gabriel Fahrenheit (1686-1736) was born in Danzig, then a semi- independent city affiliated with Poland, but German-speaking; he then moved to Denmark, and finally to Holland. He did not invent the thermometer, but did make ones that were very good for their time. The concept of expressing temperature numerically was apparently invented by the physician Galen in the 2nd century. The first thermometer (if we use the word to mean a device for measuring temperature numerically) was most likely invented by Santorio, an Italian, about 1611. Its scale, incidentally, was divided into 48 intervals each labeled as 10 minutes, with 60 minutes making up one degree. Let us be thankful that no scale with degrees and minutes of temperature has survived! Galileo also used a thermometer about this time and it almost automatically follows that he has been credited with its invention; two other names sometimes put forward are Robert Fludd or Flud of Britain (unlikely) and Cornelius Drebbel of Holland (possible, but little evidence). These thermometers used air as the working fluid and their readings were therefore influenced by the current air pressure. The first liquid-in- glass thermometer was invented by Grand Duke Ferdinand II of Tuscany, apparently sometime between 1641 and 1654. For some decades after this every thermometer had its own different scale, though sometimes two or more would be calibrated to agree with each other. The idea of defining a temperature scale by fixed points was invented by the astronomer Roemer about 1702. He used the freezing and boiling points of water as his fixed points, calling them 7.5 and 60 respectively! It was Roemer that put young Fahrenheit onto the topic of thermometers when they met in 1708. Fahrenheit began selling them in 1717. Fahrenheit published an article in 1724 in which he wrote that the graduation of his (now well-known) thermometers was based on THREE fixed points: "A mixture of ice, water, and sal-ammoniac or sea-salt", as 0, with the note "The experiment succeeds better in winter than in summer"; "Water and ice mixed together without the above-mentioned salts", as 32; and "When the thermometer is held in the mouth, or under the armpit, of a living man in good health, for long enough to acquire perfectly the heat of the body", as 96. But Middleton, considering the nonspecific phrasing of the first item and the strange note, concludes that it is in fact BOGUS: Fahrenheit was trying to disinform his competitors! This does in fact make sense. Nobody else ever used more than two fixed points to define their scale; it would be surprising if they were in a convenient numerical relationship. Notice also that 96 - 32 gives a power of 2; this means that Fahrenheit could easily draw the temperature scale on his thermometers by repeatedly halving the distance between his two fixed points. (His meteorological thermometers were calibrated from 0 to 96, and of course 32 - 0 is 1/2 of 96 - 32, also easily measured off.) Fahrenheit extrapolated his scale upwards and measured the boiling point of water as 212 degrees; "soon after his death" this was made the upper fixed point, and as thermometers became more accurate it was realized that body temperature was not quite 96 after all. (It seems to me that the .6 in the famous 98.6 is also bogus and probably arises from over-precise conversion of 37 Celsius. After all, body temperature in a healthy person is not constant to the level of a tenth of a Fahrenheit degree, right, Craig? But I've never seen this point commented on anywhere.) There then arose a great profusion of temperature scales defined in various ways. One thermometer from 1841 reads in the Old and New Florentine, Hales, Fowler, Paris, H.M.Poleni, Delisle, Fahrenheit, Reaumur, Bellani, Christin, Michaelly, Amontons, Newton, Societe Royale, De la Hire, Edinburg, and Cruquius scales. That's 18. Try THAT on your radio weather broadcasts! The Swedish astronomer Anders Celsius (1701-1744) produced the first thermometer with 100 degrees between the freezing and boiling points of water, in 1741, but it was an inverted scale with 0 at boiling. Carl von Linne, alias Linnaeus, was apparently the first to flip it right way up and produce what became known as the centigrade scale, until 1948 when it was officially renamed the Celsius scale. The other most significant scale of the 18th century was Reaumur's, named for Rene-Antoine Ferchault de Reaumur (1683-1757). He defined it in 1730 such that N degrees meant that his working fluid had 1+.001*N of the volume it had at the freezing point of water; but he used a poor working fluid and the scale was revised after his death so that 0 and 80 were the freezing and boiling points of water. Reaumur's scale was the one that centigrade had to displace in much of Europe, including France, as Fahrenheit's was in other countries. The idea of an absolute scale was invented by William Thomson, Lord Kelvin (1824-1907), in 1848. The absolute scale using Celsius-size degrees became called the Kelvin scale, and the units were renamed kelvins in recent years. Mark Brader "The singular of 'data' is not 'anecdote.'" utzoo!sq!msb -- Jeff Goldberg