62 Mr. S. R. Milner and Prof. A. P. Chattock on 



tions in the position of the slider as to render these results 

 unreliable. The apparatus might, however, very possibly 

 be improved to give a more constant temperature to the 

 water. 



The results given in Table I. are only those obtained at 

 the natural temperature of the water in the tank, and include 

 all experiments taken under this condition within a period of 

 three weeks. The various columns contain the mean experi- 

 mental data with the calculated values of the temperatures 

 and conductivities; their contents are sufficiently indicated 

 by their headings. 



The factor outside the square brackets in equation (7) gives 

 the "uncorrected conductivity " of the sixth column ; its nume- 

 rical values for the two vessels are 



R H Vessel ■ 



'o.1988xO-0476^x^x -° 4100 " 



43-73 0-1192^ 



= 2-172 xlO" 5 — . 

 z 



L. H. Vessel :— 



0-1988 x0-04769 2 8x ~-^- ? x ft1iaQ 

 4o to [) Liooz 



= 2-205 xlO- 5 ~, 



n being the number of complete turns given to the micrometer- 

 screw, z the reading on the bridge-wire, and 8 that on the 

 electrodynamometer. 



The average deviation of the corrected values of k from the 

 mean (allowing for temperature) is about l'l per cent. This 

 is almost entirely due to slight fluctuations in the flow of the 

 water through the sinks which, as already mentioned, cause 

 relatively large alterations in the position of the contact- 

 maker on the bridge-wire. There is therefore no one-way 

 tendency in these alterations, and they cannot much affect the 

 final mean. 



A comparison of the values given by the two vessels is made 

 in Table II. This is obtained by arranging the numbers in 

 Table 1. in the order of decreasing temperature, dividing 

 them into three equal groups, and taking the mean of each 

 group). 



The values given by the right-hand vessel are thus only 

 about 0'2 per cent, higher than those by the other vessel at 

 the same temperature, a result which may be easily due to 



