large for accurate readings each was divided into four, and thus arose the scale 0–48–96. When he made thermometers for higher temperatures the scale was merely lengthened by adding more spaces of equal size, and one of the divisions marked 212 accidentally coincided with the level of the liquid at the boiling-point of water; Fahrenheit never had any intention of dividing the interval between his zero and the boiling-point of water into 212 parts.
It is a singular thing, however, that if we adopt the fixed points 32 and 212, the actual temperature of the human body is 98° not 96°; so the Fahrenheit scale now in use is not exactly the original. Moreover the zero of the Florentine thermometer is given by him as equal to 45°F. and by others as 48°.
Uncertainty also exists as to the exact temperature selected by Fahrenheit for his zero; the proportions of ice, water, and salt (or sal-ammoniac) in the mixture he used are unknown. Rüdorff has shown that the temperature obtained by mixing 100 parts of snow and 33 of salt equals −21.3° C., and 100 parts snow with 25 of sal-ammoniac gives −15.4° C., whereas the modern Fahrenheit zero is equal to 7.8° C. Fahrenheit's original mixture
