168 



Mr. R. Sabine on a New Determination of 



The real resistances (Wj) of the tubes at 0° C. are therefore 

 represented by 



W 1= W + 2z/=w(l + £). 



Table V. 



No. 

 3 



5 



7 



millims. 



556-47 



19397 



1918-32 



Fig. 2. 



4 



i 



8 2602-37 



10 1541-80 



11 4382-18 



12 1652-84 



13 . . . ... 1636-82 



14 . . . . . . 1420-04 



15 1525-61 



It now remained to compare the tubes as resistances with each 

 other, in order to find in how far the calculated values would 

 then agree. 



This was done by employing successively two of the tube s as 

 the two branches x and Wj of a Wheatstone's resistance bridge, 

 the opposite branches being 

 formed by a fine tightly 

 drawn % platinum wire a 

 metre long, divided into 

 the two parts a and b by 

 the platinum contact-rol- 

 lers R. A DanielPs cell 

 and key were inserted 

 between the union (S) of 

 the two tubes and the 

 contact-rollers, an d a mirror 

 galvanoscope between the er 



ends e and / of the plati- 

 num wire. The lengths of 



a and b were read off by means of a nonius carried by the block 

 of the contact-rollers along a metre scale at the back of the plati- 

 num wire*. 



surface being introduced into the calculation as hemispherical instead of 

 plane. 



This correction was omitted in the two previous determinations as being 

 unimportant. 



* A detailed description, with a drawing of this instrument, may be found 

 in PoggendorfFs Annalen, vol. ex. p. 9. The common error, arising from 

 resistance to passage of the current from the ends of the wire to the clamps, 

 was lessened by employing a wire of nearly 20 metres resistance . The 



sc 



W. 



TTi% 



If 



R 



