SOME MERCURY STANDARDS OF RESISTANCE, ETC. 



117 



\\cir i-oastructed in 1890 and 18!K5, ami m> variation ao graftt a> 0*001 j>er 

 cent, has up to the present (1903) l)een detected. 



The whole of the observations connected with this investigation have been carried 

 out at the National Physical Laboratory. As before stated, the work was first 

 initiated by the Director in 1900, and the course followed is such as to preserve 

 continuity with the past researches of the British Association. To the Director the 

 author has to express liis great indebtedness iu ;ill departments of the work. To his 

 colleague, Mr. B. F. E. KEELING, the success of the linear measurements is entirely 

 due, while the various members of the staff are thanked for their kindly interest and 

 ,ulvice. 



APPENDIX I. 

 Explanation of the Calibration Curves and Tables in Appendix. 



All the calibration curves are plotted to approximately the same scale, the ordinate 

 represented by the side of one square l>eing equivalent to a change in cross-section of 

 very nearly 0'5 per cent. The more correct scale is given in the following table. 

 For each standard there is also given the distance from the etched mark to that end 

 of the tube nearer to the mark. This distance is required lx>th for the evaluation of 

 p! and for the calculation of the mean cross-sections of different portions of the tube. 

 See Table III. 



TABLE XVII. 



APPENDIX II. 



For each tube five values of pf were calculated from five series of olwervations, the 

 formula employed for each series being 



p.' = (2X + \., + 6X*) (21/X -f a/\ a + 6/X*)/(n + -f b)*, 

 in accordance with the text (see p. 59). 



