the Thermal Conductivity of Water. 57 



If we put k" for the thermal conductance per square 

 centimetre between heating-coil and sink 



0-94-1 



^p c 2 R = we + k'd = we' + vd r , 



from which 



*- a ca ee 



■e_l(d__d/\ 

 ' j\e &)' 



& — e was too small to be read on the piles themselves, but 

 could be measured by a direct reading on the bridge-wire. 

 The following is an example of a complete determination 

 ofF:— 



S. 6 V 6i. d x . d t '. z. 



RH. vessel... 19-4 +12-22 +12-4 +8-55 -10'32 4'1 



8 being the reading on the electrodynamometer, and z that 

 on the bridge-wire, and the suffix x being used to indicate 

 the actual galvanometer readings of the piles. Hence using 

 the constants already given 



iir x iii x (°"° 477 ) 2 x 19 " 4 x 6-15 x ( 4-1 x °' 119 ) 



0-516 x 12-22 x 12-4 x 0'588-f ^£ + ^^} 



■ =8-6 xlO" 6 . 

 The values of k r obtained were 



K.H. Vessel. 



L.H. Vessel. 



8-6 x 10- 6 



9-3 xlO- 6 



9"1 n 



8-3 „ 



8-4 „ 





giving a mean value 8*75 X 10 -6 O.G.S. 



Introducing this value into equation (7), putting & — e= 

 0*119 times the displacement (z) of the slider on the bridge- 

 wire, and writing #/, d v and d{ for galvanometer- readings 

 instead of temperatures, we obtain for the percentage cor- 

 rection due to the loss of heat upwards : 



RH. vessel -^-\ 2'55 6 ±W^A + 0-588 d^V 

 L.H. vessel -l|!|3-27 ^ ( ^~^ + 0-602rf 1 ]>. 



The average value of this correction was small, and amounted 

 only to about —0*3 per cent., since before the increase of 



