26 PHILOSOPHICAL SOCIETY OF WASHINGTON. 
whole degree. In a year after that it may rise an additional five- 
tenths of a degree; with succeeding years the *change is less and 
less. 
When a thermometer is raised to a high temperature its freezing 
point is depressed. The average depression for 100° is about 0.°2. 
On raising to a temperature lower than 100° the freezing point is 
also depressed, but not so much. For 50° the depression is about 
0.°05. For temperatures as high as 100° the depressions are about 
proportional to the squares of the temperatures by which they are 
produced. 
The cause of these changes of freezing point is in the nature of 
the glass. The mercury in the thermometer has nothing to do with 
them; neither has the atmospherig pressure. 
The amount of the changes depends on the composition of the 
glass in the thermometer-bulb. It has been recently ascertained by 
H. F. Wiebe that the change is greatest for glass containing equal 
quantities of potash and soda. A thermometer made of a variety 
of glass containing 14 per cent. of potash, 14 per cent. of soda, the 
remainder silica and oxide of lead, was found to have its freezing 
point depressed 0.°84 on raising it to 100°. Thermometers in which 
the potash or soda in the glass was replaced by lime were found 
to have the freezing points depressed only 0.°07 for the tempera- 
ture of 100°. 
To produce the maximum depression of freezing point peculiar 
to any temperature requires that the thermometer be kept at that 
temperature for a certain length of time. For a temperature of 
100° a half hour suffices; for 50° two hours are required. 
If the thermometer is kept at 100° longer than half an hour 
the depressed freezing point after that time begins to rise. If con- 
tinued at the higher temperature for two weeks the freezing point at 
the end of that time will be found to have risen about one degree. 
This fact is taken advantage of by some makers of thermometers 
to produce an instrument whose freezing point will vary but little 
in years subsequent to its manufacture. 
The depression of freezing point produced by high temperature 
is only temporary. ‘The thermometer in the course of time regains 
the reading of its freezing point corresponding to ordinary tempera- 
tures. The more quickly the depression is produced the more 
slowly the reading is regained. 
After a thermometer has been subjected to a temperature of 100° 
