102 



Ml, 1 F. K. SMITH ON THE CONSTRUCTION OF 



That the differences (a) (b) do not agree is not surprising. Apart from the 

 experimental errors, the constant /= 0795 is a mean value only. Its magnitude 

 must certainly depend to some extent on the ratio of the outer diameter of the tube 

 to that of the bore, in other words, the " flange effect." Measurement shows this 

 ratio to be approximately equal for the tubes, but the exact measurements are with- 

 out interest, since the law of variation is not known. Again, the sections are not 

 truly circular, and f is dependent on the contour. The agreement between the 

 differences (a) and (b) may therefore be regarded as sufficiently close. It will be 

 observed that the resulting sums of the differences are practically identical, from 

 which it follows that if the various values of f were evaluated from the difference 

 of the resistances in Methods I. and II., the mean f would have been identical with 

 0795 to O'l per cent. This satisfactory result is no doubt due, in a large measure, 

 to the number of standards employed. 



Method No. Ifl. Measurement of Resistance by the Carey Foster Bridge. 



In order to preserve continuity with the work of previous observers, and also to 

 better decide the most advantageous method of measuring the resistance, observations 

 were made by the Carey Foster Bridge. The method adopted was very similar to 

 that of Lord RAYLEIGH in 1882 and of Mr. GLAZEBEOOK in 1888. It is also of 

 interest to note that the standard coils and the bridge (that designed by Dr. FLEMING 

 for the British Association) were the same as employed by those observers. 



Different connections were now requisite. The standard tube was terminated 

 at each end by a glass vessel with three apertures. These are shown in the 

 figure (fig. 16). The bungs employed were treated as before, the necessary tests for 

 insulation resistance being made before proceeding with an observation. 



s s 



Fig. 16. 



Three platinum thimbles 4 centims. long and 1 - centims. in diameter at the orifice 

 were inserted in ebonite collars and fixed in position at Lj, L 2 , and L 3 as shown. 

 Rubber sheaths passed from the necks of the vessels to the outside of the ice box. 

 These sheaths were lined with glass tubes to give them greater rigidity, and also to 



