MECHANICAL EQUIVALENT OF HEAT. 129 



the latter of which I believe to be the only method of 

 eliminating constant errors. 



A story is related of a would-be physicist who carried 

 his belief in the doctrine of averages to such an extent 

 that to save himself the trouble of direct weighings he 

 adopted the custom of asking the opinion of a large 

 number of individuals as to the weight of a certain object 

 — his argument being that if he obtained a sufficient 

 number of guesses the mean result would be near enough 

 to the truth. History does not record if he adopted the 

 method of " least squares ". 



On the same principle it may be of value to give the 

 mean of all the results referred to in the above tables. 

 For observers who adopted direct methods, such as the 

 agitation of water, etc., I find it to be 431*1 kilom. per 

 thermal unit C. ; for those who adopted what may be 

 termed indirect methods, such as the velocity of the pro- 

 pagation of sound, etc., 421*0. 



Undue attention has been devoted by many observers 

 to the methods adopted for the conversion of mechanical 

 work into heat as compared with the regard paid to that 

 vital point — the thermometry. Some have even con- 

 sidered it sufficient to employ the ordinary mercury scale 

 of temperature, and the unfortunate part of such want of 

 attention is, that it is almost impossible, even with the 

 light of later investigations, to estimate the consequent 

 errors. The thermometers have in most cases disappeared, 

 and, when still available, the conditions under which they 

 were used are so uncertain that a comparison with our 

 present standards is of little avail. The reading of some of 

 the papers above referred to-is to me a painful experience. 

 The devices employed for converting nearly the whole of 

 the energy into heat, the expedients for estimating the 

 residual amount of energy converted into other forms, are 

 such as to compel the admiration of the reader, but, although 

 so much ingenuity has been directed to the adornment of 

 the upper stories, the foundation — the measurement of tem- 

 perature — has been neglected, thus rendering the whole 

 edifice unstable. 



