Chemical Affinity in terms of Electromotive Force. 387 



contraction* taking place, even when the solution is already 

 so far diluted as not to cause any measurable heat-evolution 

 on mixing with water. 



The observations were made with a large-bulbed dilatometer 

 holding about 94 cubic centim. The volumes corresponding 

 to the various graduations were determined by introducing a 

 known weight of distilled water, and determining the level of 

 the water in the stem — first, when the temperature was a little 

 below 18 0, 5, the instrument being immersed in a large vessel 

 of water kept agitated ; and secondly, when the temperature 

 was a little above 18°'5. In this way the level at 18 0, 5 was 

 obtained by interpolation. The weight of the contained water 

 (reduced to a vacuum) being known, the volume in cubic 

 centimetres occupied up to the level at 18°*5 was calculated 



W 



by the formula V = flflQgAQ ; where V is the volume in cubic 



centimetres, W the weight in vacuo, and '998602 the weight 

 of 1 cubic centimetre in grammes at 18 0, 5 (Kopp). By 

 making a series of such observations a capacity-table was 

 calculated for the instrument, expressing the volume in cubic 

 centimetres at 18°'5 for each graduation. A known weight 

 of the solution to be diluted was then introduced, and the level 

 at 18°'5 obtained by making observations a little below and a 

 little above that temperature, so as ultimately to obtain the 

 volume occupied by a given weight (reduced to vacuum) 

 of solution. Repetitions of experiments gave values accord- 

 ing together within a range of '00003 (3 milligrammes in 94 

 cubic centim.), and often within much less ; the chief source 

 of discrepancy being in the correction for reduction to a 

 vacuum which was altogether about 100 milligrammes, neces- 

 sarily varying with the temperature and barometric pressure, 

 so that a small error in temperature or pressure estimation 

 made several milligrammes difference in weight-correction. 

 On the whole, the probable error of the average of some three 

 or four repetitions of an observation was less than +'00001 

 (1 milligramme in 94 cubic centim.). 



A quantity of the solution to be diluted was then weighed 

 up in a stoppered bottle and a weighed quantity of water 

 added, the weighings being reduced to a vacuum : the whole 

 was thoroughly intermixed, and the weight of a cubic centi- 

 metre of the diluted fluid determined as before in the dilato- 



* The rough values obtaiued in § 100 (Part V.) led to the conclusion 

 that on diluting zinc- or copper-sulphate solution with water an increment 

 in volume takes place. This is now found to be incorrect, the experi- 

 mental errors of the rough observations being greater than the volume- 

 alterations to be measured. 



