1891.] On certain Ternary Alloys. 133 



their relative solubilities, i.e., two of the liquids being only 

 miscible together to limited exteuts (like lead and zinc), whilst the 

 third was miscible in all proportions with either of the others 

 separately. The difficulty of making sufficiently accurate analyses 

 of the ternary mixtures thus obtained prevented our using several 

 such groups of liquids, which at first sight suggested themselves, 

 more particularly mixtures of alcohol, -water, and ether ; but we found 

 that chloroform, water, and glacial acetic acid fulfilled all the necessary 

 conditions ; so that, when a mixture of equal weights of the first two 

 with not too large a proportion of the third was well agitated and 

 allowed to stand, it separated into two ternary solutions exactly 

 correlative with the ternary alloys previously examined ; the heavier 

 one consisting chiefly of chloroform with some of the acetic acid and 

 an amount of water proportionate to the acetic acid present ; the 

 other consisting mainly of water with the rest of the acetic acid, 

 and more or less chloroform dissolved therein. Calling any given 

 such pair of conjugate mixtures A and B respectively, we found that 

 the general principle deduced by Sir 0. G. Stokes could be verified with 

 sensible accuracy with these liquids on agitating together A and B in 

 various proportions, each liquid separated out again unchanged in each 

 case, no matter whether A was used in large excess of B, or vice versa. 

 On the other hand, when two different alloys, A and B, were made of 

 lead, tin, and zinc in such proportions that one was approximately 

 conjugate to the other as indicated by the previously recorded 

 observations, we did not succeed in getting anything like such sharp 

 results ; experiments where 2 parts of A to 1 of B were mixed 

 together, and treated side by side with a mixture of 1 part of A to 

 2 of B, did. not give quite the same results in the two cases, the 

 differences being considerably larger than anything attributable to 

 ars of analysis and such like sources of inaccuracy. 



Mixtures of Chloroform, Water, and Acetic Acid. 



The analysis of such mixtures we found could be carried out with 

 siderable accuracy and . ease in the following way ; a weighed 

 irtion of the mixture (contained in a stoppered bottle) was diluted 

 with water, and titrated with a fresh caustic soda solution accurately 

 standardised, using phenolphthale'm as indicator. Another portion, 

 eighed in a flask or bulb tube containing a. little water, was then 

 ibmitted to -the action of a current of dry air sucked through it, 

 e issuing gases and vapours being made to pass through a pumice- 

 stone and sulphuric acid drying tube. When constancy of weight was 

 attained, and all chloroform had been removed, the loss of weight of 

 e entire apparatus represented the chloroform ; whilst the gain in 

 'ht of the apparatus (as compared with its weight before intro- 



VOL. XLIX. 



s 



