Dr. E. J. Mills on Elective Attraction, 517 



Lithic Nitrate. 



This salt was prepared by treating lithic chloride with hydric 

 nitrate that had been distilled below the boiling-point_, and 

 otherwise proceeding as in the case of potassic nitrate. In order 

 to ascertain its purity, a quantity was converted into chloride 

 (the nitrate being extremely difficult to weigh for this purpose) 

 by evaporating to dryness with aqueous hydric chloride aiicl 

 fusion. The new chloride yielded 129*52 per cent, of sulphate^ 

 theory requiring 129-56; the nitrate from which it was derived 

 was free from sulphate and chloride, and was neutral. 



In order to secure as complete a desiccation as possible, the 

 nitrate, which is an excessively hygroscopic substance, was pow- 

 dered at 100^-1 80^ and rapidly introduced into the reaction-tube; 

 this with its contents was then plunged in an air-bath already 

 heated to 200°, in which state it was maintained for a period of 

 22-48 hours. Phosphoric oxychloride gave rise to immediate and 

 energetic action ; and it became turbid, with deposition of flakes 

 and crystals, at 25^-35°. At 75°-80° the liquid was again clear ; 

 at 80°-90° the principal part of the reaction took place ; at 100° 

 much (or even all) of the nitrate had dissolved, and oily streaks 

 (a sign of incipient viscidity) were observed in the mixture. On 

 gradually raising the temperature some crystals were deposited, 

 the oxychloride for the most part (but never wholly) distilled away, 

 and the mass in the tube became vesicular, with so much intu- 

 mescence as to hinder, and finally interrupt, the passage of air. 

 This last result invariably ensued, although great care was taken 

 to attain the ultimate stage very slowly, and the operation was 

 sometimes prolonged to more than five hours. 



The tube with the residue was plunged into cold v/ater, ex- 

 cepting on one occasion (experiment I.), it having been found 

 that the ordinary or converse operation was attended with danger 

 of loss owing to the frothing that ensued. Solution took place 

 quietly, and was complete in about an hour, this period being 

 required by the free oxychloride which invariably separated. As 

 the residue might contain lithic nitrate and chloride, phosphoric 

 oxide and oxychloride, and possibly some fifth constituent, it was 

 evident that the value of a would not admit of direct determina- 

 tion unless in a case where the attack was total. Only one of 

 the following experiments (V.) certainly fulfils this condition. 

 Fortunately it was found, on examination of the analytical num- 

 bers, that the ratio (r) between the weight of argentic chloride 

 and magnesic pyrophosphate was still sensibly constant. Hence 

 there is a basis of comparison between the instances of total and 

 partial attack. In experiment VII. the amount of residual ni- 

 trate was determined by Harcourt^s process, and agrees well with 



