Proceedings. 103 



method of procedure. The thermometers are kept in the 

 water for a sufficient time to allow them to acquire its 

 temperature. The standards are then read off, the carrier 

 taken out, and the thermometers may be read off 

 at leisure. The operation is repeated at the other 

 temperatures for which the testing is to be carried out. As 

 the ultimate standard of temperature, I use at the College 

 a thermometer calibrated carefully at the Bureau Interna- 

 tional des Poids et Mesures at Paris. With proper precau- 

 tions a temperature can be read off to three or four 

 thousandth of a degree Centigrade. The College also 

 possesses a thermometer divided into tenths of a degree 

 Centigrade, compared with the Standards of the Technische 

 Reichs Anstalt at Berlin. Both thermometers agree in their 

 indications; they are made of glass, the composition of 

 which in each case is definitely known, and their indications 

 may without trouble be reduced to the air thermometer if 

 desired. 



" I have, further, two thermometers compared at Kew. 

 One of them (A) is divided into fifths of a degree Fahren- 

 heit, and ranges from 90 to 115. Unfortunately it does not 

 contain the freezing point, so that its changes cannot be 

 followed. The other (B) is divided into tenths of degrees, 

 and ranges from 90 to no. Another part of the stem, 

 separated from the rest by a bulb, is divided from 30 to 

 36 , so that its freezing point may be tested at any time. 

 Finally, three clinical thermometers are also used for com- 

 parison, two of them having been standardised at Berlin 

 and one at Kew. 



"There is some doubt as to what the scale of temperature 

 used at Kew really is ; but the difference between the Kew 

 temperatures, and the scale used on the continent, being 

 probably about ^th Fahrenheit near ioo° F., is of no im- 

 portance as regards clinical thermometers. 



"The following comparison shows the agreement between 



