168 THE REALITIES OF MODERN SCIENCE 



to obtain by the best means at his disposal was the 

 lowest obtainable by any means and thus take that 

 temperature as the zero of his scale, as did Fahrenheit, 

 or (2) he might arbitrarily assume a zero. The latter 

 course was followed by Celsius, who in 1742 devised 

 what we know to-day as the " Centigrade scale. 7 ' 



The zero chosen by Fahrenheit corresponded to the 

 temperature reached by a mixture of sal ammoniac 

 (ammonium chloride) and melting snow. In the 

 Centigrade scale the zero is the temperature of melting 

 ice and water. As a matter of fact it was Fahrenheit 

 himself who had made this selection possible by show- 

 ing, sometime previous to Celsius' choice, that the 

 temperature of such a mixture is constant as long as 

 the ice is not entirely melted. The difficulty previous 

 to Fahrenheit's demonstration had been that it was 

 known that water could be cooled below this temper- 

 ature without ice forming. This is true, but the con- 

 dition is essentially unstable, for if a small bit of ice 

 is dropped into the water freezing occurs with great 

 rapidity. 



As an upper point on his scale Fahrenheit chose, 

 unfortunately, the blood temperature of the human 

 body. He then divided the interval into 96 equal 

 divisions. Celsius, however, selected the boiling point 

 of water at atmospheric pressure and divided the 

 temperature interval between zero and this upper value 

 into 100 equal degrees. 



The positions of the liquid in the stem of a ther- 

 mometer like that of Galileo, corresponding to the two 

 arbitrarily assumed but easily reproducible tempera- 

 tures of the Centigrade scale, may be marked and the 



