456 



Mr. J. D. Hamilton Dickson on 



were constructed for eight different and very carefully made 

 platinum-thermometers. There does not seem sufficient cause 

 to choose only the five points, especially when by the method 

 of least squares each of the eight observations could have 

 been made to contribute its quota towards getting that rela- 

 tion which, by all mathematical methods at our disposal, is 

 nearest to the truth as displayed by the observations. 



Of the eight observations given in (Gr.) I am unable to 

 find data to enable me to use methyl salicylate. From the 

 other seven as given in tables I. and x. (Gr.) for thermo- 

 meter G (taking the mean of the two values of the boiling- 

 point of mercury in table I.), I have, by least squares, 

 obtained the formula 



(R, + 22-897523) 2 =-91924237(* + 843-3048). . (15) 



In the first line of Table V., after the first three headings, 

 appear the names of the five substances used by Mr. Griffiths, 

 with the respective boiling-points he adopted placed below 

 them in the 2nd line. The 3rd line contains the tempera- 



Table V. 



Absolute 

 zero. 



Ice. 



Steam. 



Aniline. 



Naph- 

 thalene. 



Benzo- 

 phenone. 



Mercury. 



Sulphur. 





o 

 



100 



184-41 



218-06 



300 08 



35772 



44838 



-272-95 



0-57 



99 39 



18374 



217-87 



306 80 



358-55 



447 73 





-•57 



4-61 



+ ■67 



+ •19 



-•72 



-•83 



+ •65 



-27245 



0-45 



99-25 



183-66 



217-82 



300-88 



358-74 



448-11 





-•45 



+ •75 



+•75 



+ •24 



-•80 



-1-02 



+ •27 



-278-66 



044 



99-66 



183 85 



217-79 



305-94 



357-06 



44486 





-•44 



+ •34 



+•56 



+ ■27 



+•14 



+ ■66 



-33 



Note. — Numbers in italics were not employed in calculating the corresponding 

 formula?, but these residts are entered in order to complete the table. 



tures calculated by (15) from the resistances in Mr. Griffiths's 

 table x. ; and the 4th line shows the divergences between 

 the 2nd and 3rd lines, from which the probable value of a 

 divergence, calculated as before, is o- 43. The absolute zero 

 (R=0) given by this formula is — 272°-95. 



I also used Mr. Griffiths's five fixed points from which to 

 develop a formula like the above, namely 



(R + 22-4407292) 2 = -90473803(£ + 829-05810). . (16) 

 The 5th line of Table V. contains the temperatures calculated 



