METHODS EMPLOYED IN CALIBEATION OF MERCURIAL THERMOMETERS. 145 



Report of the Committee, consisting of Professor Balfour Stewart, 

 Professor EliCKER (Secretary), ami Professor T. E. Thorpe, ap- 

 poliited for the purpose of reporting on the Methods employed in 

 the Galihration of Mercurial Thermometers. 



[Plates I., IL, III.] 



Introduction. 



In drawing up tlie fcllowing Report the Committee have desired to pre- 

 sent it in the form in which it will be most generally useful. It is there- 

 fore divided into two parts. The first contains a brief description of the 

 principal methods of calibration and correction which have been hitherto 

 proposed, an account of the thermometers on which these methods have 

 been tested by the Committee and of the apparatus employed, and a 

 summary of the results arrived at. This portion therefore contains the 

 facts necessaiy to enable a selection from the various methods to be made 

 by persons intending to undertake the calibration or correction of a 

 thermometer. The second part consists of a fully worked out example of 

 each of the methods, together with remarks of a detailed character. This 

 part will it is hoped be useful in facilitating the calculations required, 

 especially as references to the subject of calibration in English scientific 

 works are rare and meagre. 



Part I. 

 General Beviern of Methods of Calibration ami Correction. 



(1) The corrections for the inequalities in the bore of a thermometer 

 tube may be applied in two different ways, distinguished as methods of 

 calibration and correction respectively. 



In the first the length of a column of mercury is measured in various 

 parts of the tube before the scale is etched on it, and the lengths of the 

 divisions are then so adjusted as to make equal differences of scale- 

 readings correspond to equal volumes. In the second the tube is in the 

 first instance furnished with a uniform scale, and a table of corrections 

 is afterwards drawn up, by means of which the same end is attained as 

 before. It will be shown hereafter that a high degree of excellence can 

 be attained by the former method, but although generally used, the 

 scales, even on good thermometers by well-known makers, are not, as 

 a rule, sufficiently correct for \evj accurate work. They therefore 

 require collection, and the process is rendered very laborious by the 

 varying lengths of the divisions which have to be measui-ed and allowed 

 for in the calculations. The results, too, are probably less accurate than 

 they otherv.'ise would be on account of the irregular forms of the correc- 

 tion curves produced by the superposition of the errors of the tube and 

 scale. A uniform scale, the divisions of which are about a millimetre 

 apart, seems therefore in general the most convenient. Such tubes only 

 should be used as preliminary tests have shown to be of fairly uniform 

 bore. 



(2) As all the methods of calibration and correction to be described 

 1882. L 



