496 Sir W. Ramsay and Dr. B. D. Steele on the 



crystallization at very low temperatures has, so far as we are 

 aware, not been described, a short account of it may prove 

 interesting. 



Some ordinary crude toluene was cooled by pouring liquid 

 air into it until it was partially solid. The purified toluene 

 was introduced into a tube into which was sealed an inverted 

 filter, plugged with cotton-wool. On dipping this tube into 

 the cooled toluene the pure toluene slowly froze, forming- 

 large crystals ; when about one-third had frozen the un- 

 frozen portion, containing any possible impurities, was filtered 

 off. The crystals when melted formed the sample termed B. 



(6) Ether. — Prepared from absolute alcohol in the usual 

 way. It was then fractionated from phosphorus pentoxide 

 until the boiling-point was constant. The samples of ether 

 were treated in three different ways ; sample A was distilled 

 from phosphoric anhydride into the bottle in which it was 

 preserved ; sample B was a portion of A which had been 

 allowed to stand for three days over phosphoric anhydride, 

 and then shaken with mercury to remove possible ethyl 

 peroxide ; it was then distilled into the weighing-bulb 

 from pentoxide by a method to be described. As samples A 

 and B both gave the same puzzling density, and as it was 

 possible that the methods of purification left something to be 

 desired, sample C was a portion of B which was frozen in a 

 manner similar to that described for toluene. It was cooled 

 in a tube immersed in frozen crude ether until the ether in 

 it had crystallized. The large clear crystals, some of which 

 were two inches long, were separated from twice their volume 

 of mother-liquor by upward filtration. The crystals when 

 melted formed sample 0. 



(7) Methyl alcohol. — Methyl oxalate was prepared from 

 " pure '' methyl alcohol. The crystallized oxalate was washed 

 with water until the washings gave no iodoform reaction. 

 The ester was decomposed with potash, and the alcohol dried 

 with lime of A^ery high quality, which reacted violently and 

 at once with water. The sample was re-distilled with lime 

 until the boiling-point became constant. 



It is constancy of boiling-point, and not temperature, which 

 we regard as a criterion of purity. To register a really ac- 

 curate temperature is a difficult undertaking, and one which 

 was unnecessary for our purpose. 



Temperatures of Experiment. 



These were approximately 100°, 115°, and 130°. They 

 were obtained by surrounding the tubes to be heated with 

 the vapour of pure chlorobenzene, which had been very 



