Thermometers at the Kew Observatory. 227 



English physicists are dependent for their accuracy upon that 

 of the verifications at Kew. Many thousands of thermometers 

 have ah-eady been \'erified by the apparatus about to be described. 



Up to the year 1875 the apparatus for this purpose at the 

 Kew Observatory was of the rudest character. 



It was simply a glass jar 9g inches wide and 18 inches deep, 

 filled with hot water and standing on a turntable, in which a brass 

 frame was placed. 



The thermometers were attached to this framework, and the 

 observer having well agitated the water with a plunger, read the 

 instruments in succession through the glass as he turned the jar 

 round before him, reading each thermometer as it passed. He 

 first turned it round from right to left, and then back again from 

 left to right. Each thermometer was thus read twice, and the 

 mean of the pair of readings was taken. It is obvious that if the 

 rate of cooling of the water be uniform, and if the thermometers 

 are observed at precisely equal intervals, the mean of every pair of 

 observations would be strictly referable to the temperature of the 

 water at the same moment of time, namely to that which is half- 

 way between the beginning and end of the entire set. It is need- 

 less to point out that these conditions can never be strictly ful- 

 filled, although, notwithstanding the imperfection of the process and 

 the coarseness of the apparatus, the observers acquired much 

 certainty and skill in its manipulation. Still the time occupied was 

 unnecessarily great, and the chance of error, owing to variations 

 in the rate of cooling of the water, was larger than it need be. 

 Partly owing to this latter reason, and partly to the fact that the 

 number of thermometers sent to be tested has considerably in- 

 creased (being now not less than 3000 annually), I thought it 

 advisable to design and propose to my colleagues of the Kew Com- 

 mittee the construction of an instrument of a much more sub- 

 stantial and adequate character; and to this the Committee assented. 

 I was subsequently indebted for many suggestions to Mr. De La 

 Eue, and also to Mr. E. Munro, of 24 Clerkenwell G-reen, London, 

 by whom it was finally made. It has now been at work for two 

 years, and its performance is quite satisfactory ; experience has 

 in the mean time suggested a few emendations and simplifications, 

 and I will therefore describe the instrument as at present in use. 



The apparatus (see figs. 1 & 2) consists essentially of four 

 parts : — 



(1) A water-vessel. 



(2) An agitator, worked by a handle on the outside. 



(3) An external heating arrangement. 



(4) A frame on which to hang the thermometers, turned 

 by a handle on the outside. 



(1) The Water-vessel. 



This is a cylinder of stout copper, 2 ft. 2 in. high and 1 ft. in 



diameter. In its base there is a central aperture through which 



the concentric vertical axes are passed, which respectively carry 



the agitator and the thermometer frame ; the top of the cylinder 



Q2 



