Platinum Temperatures. 449 



*te= R "^277 57 ° X 1Q0 = 107 ' 8R -329-52, 

 ^f= R T^ 6 o 62 xlOO= 66-43R-316-63, )- . (8) 

 ^G= R 7t' Q 9 ^ 43 xl00= 63-07 R-312-46. 



Assuming that the absolute zero is where R vanishes, these 

 formula? give for it the respective platinum-temperatures 

 -329-52, -316-63, -312-46. Applying the corrections 

 deduced by (6) and table xn. (Gr.), these become on the 

 normal air-thermometer - 245°-99, — 253°-06, — 2o9°50. 

 This formula is therefore not trustworthy for extrapolation. 



For each of the formulas suggested (I have passed over 

 those which have been already rejected by others) there 

 appears therefore to be some objection not easily overborne ; 

 but naturally there are close agreements, otherwise they would 

 not have been proposed. Probably the direction in which 

 investigation has been made has tended to the expression of 

 the resistance in terms of the temperature ; but when the 

 problem is reversed and we seek to determine temperature 

 by resistance it is equally natural to write 



t = a / + b'n + c'R 2 + ..., 



where a' , b', c'. . . are constants. 



These and other circumstances have led to the considera- 

 tion of a formula of the form 



(n+ a y=p(t+b), (9) 



where a, p, and b are constants, as being more representative 

 of the connexion between temperature and resistance than 

 any formula hitherto proposed, and at least as simple as any. 

 It evades many, if not all, of the difficulties appearing in 

 the formulae already employed. Beginning at the zero value 

 of R, t and R increase together indefinitely ; there is no 

 maximum value for R beyond which it cannot increase, and 

 there is always one and only one value of t for each value 

 of R. [It may be said, however, that there are two values of 

 R for each value of t. The answer is doubly against this 

 contention ; (1) such a value of R would be negative, which 

 is inconceivable ; and (2) even if a negative value of R were 

 conceivable, the double values would only arise below the 

 absolute zero, which is again inconceivable.] As to its posi- 

 tive merits, it, too, has three disposable constants, like (2), 

 (3, 4), and (5) ; and if it is still thought desirable to retain 



