448 Mr. J. D. Hamilton Dickson on 



not unfair to believe that 6*576 R might have been within 

 the range of attainable values in their experiments had not 

 the apparatus unfortunately broken down. 



On p. 139 of C. and Gr. data are given by which 8-formulse 

 (4) may be calculated for three thermometers named E, M l5 

 and M 2 . There are three observations for E, six for M l5 and 

 five for M 2 . Making use of these we get the formulas, 



OTE = 108*155R-331*149 ; f-*r E =1-7564 U^)~ -j^} 



» Ml = 69-315R-295-973; t-«* =1-56394 { (^f - ~] } (7) 



«r M = 69*793R-295*750; «-«r Mi =l-56447 {(j^f ~ j^ } 



whence the platinum-temperatures of the absolute zero 

 (R = 0) are -331*149, -295*973, -295*750; or, employ- 

 ing the corrections given by the 8-formulse of (7) (namely 

 + 25*077, +18*329," +18*311), they become -306°*072, 

 — 277°*644, and — 277°*439 on the normal air-thermometer. 



To formula (5), for which Professor Callendar found 

 (though only as an " odd coincidence," Call. p. 196) that the 

 values a = *003,425 } 9, /3 = '001,529 "agree best with the 

 observations" (Call. p. 190), there are serious objections. 

 According to it, R vanishes only when 1 + fit is positive 

 and infinitely small, that is, when t has a negative value 

 indefinitely little less numerically than 1//3 or 654*022, 

 — in other words, R vanishes indefinitely near the (quasi) 

 temperature of — 654°*022 but above it, while immediately 

 below it R is infinite. On the other hand, there is an asym- 

 ptotic value for R as t increases to infinity, namely Roe a//3 , or 

 about 9*4 R . In the experiments of Professors Dewar and 

 Fleming we have evidence that the resistances of all pure 

 metals tend to zero at or near the absolute zero ; and there 

 is no reason at present known why there should be an upper 

 limit to the resistance with the increase of temperature. 

 Hence this formula is generally inadmissible. 



Mr. Griffiths' s formula (6), having four disposable constants, 

 is, so far, better than one with a less number, throughout the 

 range over which it is used, namely from 0° to 400° C. But 

 again the objections to it are not easily overcome. In table xn. 

 on p. 59 of (Gr.) he gives the values of a, b, c, d for three 

 thermometers named E, F, G ; and on pp. 65, 67, and 68 he 

 gives w for each of these thermometers, namely 



