Dr. Mills's Researches on Thermometry. 187 



self supposed — that he has removed, instead of imposing a 

 burden. If he had bestowed some of the attention he has de- 

 voted to the variations of the correction-factor on those of the 

 correction itself, he would not have left it for us to point out 

 that the latter are negligible. As it is, the error introduced 

 by taking the mean correction for similar thermometers instead 

 of that for the thermometer actually employed, never exceeded 

 the probable error of an observation ; and the chances appear 

 to be three to one that, with a given thermometer, under the 

 most unfavourable circumstances of exposure and temperature- 

 difference, it cannot exceed half that amount. 



Passing on to the next point noticed by Dr. Mills, he is 

 mistaken in thinking that, by a confusion between the correc- 

 tion-factor and the correction, we have attributed to him the 

 statement that it is possible " for the correction for a thermo- 

 meter with an exposed column 166 divisions long to be equal 

 to that of another when no part of the column is exposed." 

 As a matter of fact his paper does, owing to a misprint, impli- 

 citly contain this statement; and the passage to which he refers 

 is a reductio ad absiirdum to show that a misprint must exist. 



Zero-movements. — On the question of nomenclature we 

 have little to add. We agree that words are sometimes used 

 in other than their usual significations in "physical writings/' 

 In such cases, however, good reasons for departure from ordi- 

 nary usage are generally forthcoming, and the word is defined 

 on its first introduction in its new sense. 



We now come to the point which Dr. Mills fails to under- 

 stand ; and in our further remarks upon it we will confine our- 

 selves to the case of Henrici's thermometer. 



Unfortunately, we cannot discuss the question " mathema- 

 tical formulae apart," as Dr. Mills seems to wish. The whole 

 point turns upon the deductions drawn by Dr. Mills from 

 his mathematical formulae, which cannot, therefore, be dis- 

 pensed with. 



The "total remaining ascent" of Henrici's thermometer is 

 expressed by Dr. Mills by the equation 



y=2-100(-931) a3 --099(l-360) aj . . . . (1) 



We concluded that the position of zero (Z) is connected with 

 y by the relation 



z—2+y, (2) 



a conclusion supported by the fact that this equation holds for 

 all the five values of y and of the " zero observed," given in 

 Table VII. But if this is so, since y has no maximum or 

 minimum value, Z cannot have one either. The motion of the 



