132 DE. B. D. STEELE, DR. D. McINTOSH AND DR. E. H. ARCHIBALD 



As a general rule, the same substances conduct better when dissolved in hydrogen 

 chloride than in the other solvents. Next, in the order given, come solutions in 

 hydrogen bromide, iodide, and sulphide. An exception to this rule is found in the 

 case of ether, which conducts best in hydrogen chloride, and worst in hydrogen 



bromide. 



The conductivity temperature coefficients do not appear to have anything in 

 common with the viscosity temperature coefficient, so that the ions cannot be looked 

 upon as being surrounded with an atmosphere of the solvent (KOHLRAUSCH, ' Roy. 

 Soc. Proc.,' 71, 338, 1903). 



The coefficients are for the most part positive, the conductivity increasing with rise 

 of temperature, an exception occurring in the case of hydrocyanic acid dissolved in 

 hydrochloric acid. 



The Determination of Molecular Weights. 



The molecular weights were determined by measuring the rise in boiling-point 

 which was brought about by the addition of known quantities of the dissolved 

 substance. 



Considerable experimental difficulty has been experienced during the progress of 

 the work, which has also proved expensive on account of the very large quantities of 

 carbon dioxide which were required, and consequently only a few determinations have 

 been made. 



The accurate measurement of small differences of temperature at low temperatures 

 has been successfully accomplished by the use of a differential method, in which two 

 platinum resistance thermometers were employed to measure the temperatures, one of 

 the thermometers being immersed in the pure boiling solvent and the other in the 

 boiling solution.* 



The thermometers were each made from about 2 metres of 6 mil wire, and had 

 exactly the same resistance, which was of such a magnitude that a difference of 1 in 

 the temperature of the two coils produced a displacement of 16 '7 millims. in the 

 balance point on the bridge. The thermometers were supplied with compensation 

 leads in the usual way. The bridge was of the Carey-Foster type. With the 

 galvanometer used a difference in temperature of 0'03 C could be detected with 

 certainty. 



The apparatus was tested by immersing the two coils in (a) boiling water, 

 (b) melting ice, (c) boiling hydrogen sulphide, (d) boiling hydrogen bromide, and 

 (e) boiling hydrogen chloride ; the same balance point was obtained in each case. 



The measurements were made in the two pieces of apparatus shown in fig. 8, 



f The bridge and resistance thermometers were lent to us by Dr. H. L. BARNES, through whose advice 

 and assistance many difficulties have been avoided. We take this opportunity of expressing our thanks to 

 him for his kindness and help. 



