374 MESSRS. R. T. GLAZEBROOK AND T. C. FITZPATRICK 



the value of the ohm expressed in centimetres of mercury at does not differ from 

 IOG'31 by more than '02 of a centimetre, or two in ten thousand. 



From this result the values found by LORENZ and Lord RAYLEIGH differ appre- 

 ciably. With regard to LORENZ'S value, we may notice that the comparison between 

 his tubes and the B.A. unit was very far from being direct. 



/ 



The tubes were compared with a Siemens' unit issued by SIEMENS and HALSKE, and 

 this with a copy of the B.A. units sent to LORENZ by Lord RAYLEIGH. The tempera- 

 ture coefficients of the two coils and of the mercury required to be known, and 

 corrections introduced. The final value found by LORENZ for the ohm in centimetres 

 of mercury is 105 '93. 



In this determination, tubes were used 1 metre long, and 1 , 2, and 3 centimetres in 

 diameter, and some part of the large difference may possibly be due to the fact that 

 the lines of flow of the current near the ends of the tubes can hardly have been 

 cylindrical. 



No such explanation, however, can be offered of the difference between Lord 

 RAYLEIGH'S result and our own. His comparisons were direct, and the results of 

 the observations on the various tubes employed are extremely concordant ; the tubes 

 actually used by him have since been broken, but the end pieces are still at the 

 Laboratory. We thought it was worth while to fit up one of our tubes, No. VI., with 

 his end pieces, and find its resistance. In this way, we were able to eliminate any 

 error which might have occurred in the resistance of connexions, as new connexions 

 were made for the purpose and had their resistance specially determined. 



The experiment was made on May 19th, and the value found, when the temperature 

 in the terminals was about 3, was 



1 -00000 B. A. unit. 



The mean value for VI. previously found, the temperature in the end pieces being 

 l c '4 C., was 



99996 B.A. unit. 



Thus, this experiment fully confirms the value we had already used, and shows that 

 no error can have been introduced by the connexions. Lord RAYLEIGH has himself 

 pointed out that the fact that the temperature of the mercury in his terminal cups 

 was from 5 to 6 C. would lead to an over-estimate of the value of r, and he con- 

 cludes that this over-estimate may in his case have been as much as '00008. 



In the course of our observations we had several times determined the resistance of 

 a tube as the mercury in the end pieces cooled down from 9 or 10, the temperature 

 of the room, to the temperature at which it was finally steady, which was on the 

 average about 1'4 C. 



Table XI. gives the results of these determinations for VI. Each horizontal line 

 refers to the same filling. Taking the resistance when the mercury in the end pieces 

 was at 2 as 1 B.A. unit, the Table gives the temperatures at which the resistances 

 were measui-ed, and the increase of the resistance of the tube up to the temperature 

 in question. 



