106 PROF. F. CRACE-CALVERT ON THE 



appeared to me an easy matter to solve, still I have had 

 many difficulties to overcome, which will be described as 

 they occurred in the course of my investigations. 



The first series of experiments consisted in placing cri- 

 noline and steel wires, very carefully cleaned, in tubes con- 

 taining pure dry or moist oxygen ; and in another series 

 of tubes, containing the same gas, was added I per cent, 

 of carbonic acid or a trace of ammonia. 



Before describing the facts observed, I must state that 

 the oxygen was prepared from pure chlorate of potash, 

 mixed with a little bin oxide of manganese, and the gas 

 obtained passed through several feet of U-tubes filled with 

 glass moistened with sulphuric acid and caustic potash. 

 The carbonic acid was purified by washing it with water 

 and passing it through U-tubes containing sulphuric acid. 

 The ammonia was purified by first passing it through a 

 saturated solution of ammonia and then over caustic lime. 

 These purified gases were then introduced into tubes having 

 i centim. diameter, and 30 centims. in length, which had 

 been previously filled with dry mercury, and containing an 

 iron blade, to the bottom of which had been fixed a small 

 mass of gutta percha, so as to isolate the iron from the 

 mercury, and prevent galvanic currents. 



The results of the above experiments were so unsatis- 

 factory that I was led to infer that I had overlooked some 

 source of error ; and it was so : small globules of mercury 

 having adhered to the perfectly polished surfaces of the 

 iron, had become centres of galvanic action — determining 

 the oxidation on the blades, and thus rendering the series 

 of experiments discordant. 



To overcome this source of error, cleaned blades of steel 

 and iron, having a gutta-percha mass at one end, were in- 

 troduced into similar tubes to those employed in the pre- 

 vious experiments, and were placed over a mercury-trough ; 

 the atmospheric air was displaced by a current of pure oxy- 



