Kew Observatory. 321 



curing at will an artificial atmosphere, which was accomplished 

 by the successful construction of a receiver, with plate-glass 

 windows, into which an additional inch of air might be intro- 

 duced, or from which three inches might be abstracted. The 

 comparison might thus be made between 31 and 27 inches, 

 a range which comprehends all weathers. 



In the next place, with regard to thermometers, the com- 

 mittee undertook to supply all Fellows of the Eoyal Society, 

 and members of the British Association who chose to incur 

 the necessary expenditure, with standards of their own con- 

 struction, and such were likewise supplied to the leading opti- 

 cians, becoming in their hands, as it were, the parents of a 

 host of accurate thermometers. Nor did the labours of the 

 committee end here, for besides thus indirectly supplying the 

 public with a better description of instrument, great facilities 

 were afforded for the verification of all thermometers which 

 might be sent to Kew. By way of variety, let us here give a 

 short sketch of the method employed in constructing a standard 

 thermometer. Our readers are well aware that in every such 

 instrument there are two points which must be accurately 

 determined before graduation, the first of these being the 

 melting point of ice, and the second the boiling point of 

 water. 



Let snow or pounded ice be put into a wooden box, and 

 left for some time in a room, the temperature of which is about 

 32°, and further let the water which forms be allowed to drain 

 off through a few small holes in the bottom of the box. Now 

 introduce your thermometer tube, which had better be an old 

 one, into the mixture, and when it has remained there for some 

 time, make a mark on the tube at the termination of the 

 mercury. This point must denote 32° if you intend making a 

 Fahrenheit thermometer. 



But if the melting point of ice be constant, not so the 

 boiling point of water. Were the pressure of the air to fall to 

 29 inches, water would boil at 210 V ; were it to rise to 30£ 

 inches, the same fluid would boil at 213°. The barometer 

 must, therefore, be consulted when the upper point of tho tube 

 is marked off, and not only must tho bulb, but also tho whole 

 column of the instrument up to the termination of the mercury 

 be immersed during the operation in boiling water, or what is 

 better still in the steam which escapes from it into the air. 



When you have thus obtained your two points, say 32* 

 and 212° of the capillary bore of your tube bo constant 

 throughout, you have only to divide the distanco between 

 these into 180 equal parts. But if the bore, as is always tho 

 case, be not uniform, you must make your degreo longer when 

 it is narrow, and shorter where it is wide ; you, therefore, 



