Intelligence and Miscellaneous Articles. 553 



composition, either through the distillation itself, or by some in- 

 fluences accompanying it. These circumstances were,— 1st. A tarry 

 residue, from a liquid which when introduced into the retort was 

 perfectly colourless. 2nd. The formation of a small quantity of 

 water in the commencement of such a distillation, though none was 

 contained in the substance used. 3rd. The gradual lowering of the 

 boiling-point of the whole liquid by a great number of distillations. 

 These facts, taken in conjunction, naturally suggested that the oxy- 

 gen of the air contained in the retort might act upon the substance, 

 and thus gradually reduce it to hydrate of phenyle. 



In order to test the correctness of this hypothesis, the atmospheric 

 air was expelled from the distilling apparatus by dry hydrogen gas, 

 and the distillation performed in a pure atmosphere of this gas. A 

 great number of distillations performed in this manner were at ex- 

 actly the same temperature, and all the other anomalies were simul- 

 taneously removed. It was however found that the liquid always 

 boiled at a lower temperature in hydrogen than in atmospheric air, 

 the difference being about 2° Cent., and this without any alteration 

 of the pressure on the surface of the boiling liquid. A similar fact 

 was noticed in the distillation of hydrate of phenyle, and also of some 

 other liquids. 



LXVI. Intelligence and Miscellaneous Articles. 



ON A NEW METHOD OY ALKALIMETRY. 

 BY ASTLEY PASTON PRICE^ PH.D., F.C.S. 



HAVING had occasion, some time since, to test the comparative 

 value of the ordinary alkalimetrical processes, I was somewhat 

 surprised at the discrepancies in the results I obtained, and also at the 

 somewhat difficult and tedious manipulation necessary, when com- 

 paratively accurate determinations were required. The most serious 

 impediment in carrying out the usual alkalimetrical processes, arises, 

 as is well known, from the liberation of carbonic acid, the presence 

 of which, even in exceedingly small quantity, being sufficient to mask 

 the point of saturation, and to prevent the indication of the presence 

 of either an excess of acid or of alkali. 



After having experienced the difficulties attendant on its presence, 

 and the almost impossibility of rapidly and entirely expelling the 

 liberated carbonic acid from solution, it appeared to me most 

 desirable to seek an alkalimetrical process, in which the carbonic 

 acid should be expelled previous to determining the saturating power 

 of the alkali under examination. It further appeared to me, that in 

 the alkalimetrical processes now in use, there existed another very 

 serious inconvenience, in that the per-centage of alkali was deter- 

 mined directly, and not indirectly ; that is, that the alkali was esti- 

 mated, and not the impurities which might be therein contained. 

 This may perhaps be better understood by an example. Pure car- 

 bonate of soda contains about 58*5 per cent, of alkali, but com- 



Phil Mag. S. 4. No. 55. Suppl Vol. 8. 2 



