152 



Mr. T. Andrews. 



[May b 1 , 



is small, and it would be easy to detect the effect of the above pro- 

 portion, or of less than one ten-thousand-millionth of a grain of 

 chlorine in one-tenth of a cubic centimetre of distilled water by this 

 process. The same kind of action occurs with other electrolytes, but 

 requires larger proportions of dissolved substance. 



As the degree of sensitiveness of the method appears extreme, I 

 add the following remarks : — The original solution of washed chlorine 

 in distilled water was prepared in a dark place by the usual method 

 from hydrochloric acid and manganic oxide, and was kept in an 

 opaque, well-stoppered bottle in the dark. The strength of this liquid 

 was found by means of volumetric analysis with a standard solution 

 of argentic nitrate in the usual manner, the accuracy of the silver 

 solution being proved by means of a known weight of pure chloride 

 of sodium. The chlorine liquid contained 2*3 milligrammes, or 

 0*03565 grain of chlorine per cubic centimetre, and was just about 

 three-fourths saturated. 



One-tenth of a cubic centimetre of this solution ("No. 1"), or 

 0*003565 grain of chlorine was added to 9*9 c.c. of distilled water and 

 mixed. One cubic centimetre of this second liquid (" No. 2 "), or 

 0*0003565 grain of chlorine was added to 99 c.c. of water and mixed ; 

 the resulting liquid ("No. 3") contained 0*000003565 grain of chlorine 

 per cubic centimetre. To make the solution ("No. 4") for exciting 

 the voltaic couple, successive portions of one-tenth or one-twentieth 

 cubic centimetre of " No. 3 " liquid were added to 900 c.c. of distilled 

 water and mixed. 



I have employed the foregoing method for examining the states 

 and degrees of combination of substances dissolved in electrolytes, 

 and am also investigating its various relations. 



IV. " Electro-chemical Effects on Magnetising Iron. Part II."* 

 By Thomas Andrews, F.R.S.E., F.C.S. Communicated by 

 Professor G. G. Stokes, P.R.S. Received April 9, 1888. 



The novel electro-chemical effects observed between a magnetised 

 and an unmagnetised bar when in circuit in certain solutions, recorded 

 in the first part of this research, were of such an interesting character 

 that I thought it desirable to extend the investigation. The present 

 memoir contains the results of a further study of these magneto- 

 chemical phenomena, which were found to be more marked and 

 characteristic when experimenting with some of the reagents herein 

 described. The method of experimentation was generally similar to 

 that pursued and described in Part I, though it was necessary to 



* For first part see vol. 42, p. 459. 



