must have contained the two salts in proportions now indeterminable.
According to Boerhaave, Fahrenheit's zero coincides with the greatest natural cold observed in Iceland in the winter of 1709, and this is sometimes stated to have been the origin of the lower fixed point in the scale. Surely that winter was remarkably mild in frozen Iceland, for zero is often exceeded in countries not regarded as arctic; yet Boerhaave remarks that "nature never produced a cold beyond zero." The Meteorological Yearbook for Denmark shows that the temperature fell in March 1888, at Stykkisholm, Iceland, to −8.5° F.
There is another element of uncertainty in Fahrenheit's scale. Both Musschenbroek and Boerhaave state that the bulb of Fahrenheit's thermometer contains 11,124 parts of mercury at zero and that when the bulb is placed in melting ice the metal expands 32 of these parts; but Boerhaave, in another place, says the bulb contains 10,872 parts of mercury, and in still a third passage he gives the number of parts as 11,520, which Dr. Martine apprehends is nearer the truth. These vaguely named "parts" depend upon the figures taken for the expansion of mercury; to show their
