334 Transactions of the Canadian Institute. [Vol. VII. 



As indicated in these tables there are some rather peculiar 

 phenomena, some of which are not easy to explain. Experiment I. a. 

 was simply a continuation of I., as the specimens were weighed and 

 returned to the jars without being cleaned of the accumulated rust. 

 The same is true of II. a. 



In both cases where the specimens were put in, well-cleaned of rust, 

 those in the atmosphere under the influence of sea-water had made the 

 largest proportional increase in weight, showing that the sea-water 

 affected the atmosphere, which in turn caused an addition in weight of 

 the iron due to oxidation. In jar No. 2, in which was the salt solution, 

 there was found a greater percentage increase than in that of the 

 distilled water. This, it was expected, would show that salt and water 

 would always produce an increase in weight over that of the distilled 

 water, and that the substances in solution had to do with the increased 

 accumulation of rust, but in all the subsequent weighings this was 

 reversed, — that of the salt and water showing little or no increase in 

 weight. This seemed to upset any preconceived notions, or any 

 conclusions that might be drawn from the first experiment, but when 

 one takes into consideration other atmospheric conditions there is an 

 explanation which may be fairly reasonable. This peculiar phenomenon, 

 as was mentioned before, was associated with the difference in 

 atmospheric humidity in the room where the experiments were 

 performed. Almost all through the course of the first experiment the 

 humidity was from seventy to eighty, and temperature fifty-five to sixty 

 degrees F., while during all the other experiments it was at thirty-five 

 to forty and temperature sixty-five to seventy. At first, when the 

 humidity was about seventy-five, the jars remained moist for about 

 twenty to twenty-four hours, whereas in the later experiments the 

 moisture disappeared in five to six hours in jars two and three. 



At first it was thought sodium chloride and other salts jnight pass 

 into the air, but these experiments point rather to another conclusion, 

 namely, that the salts affected the formation of iron oxide, only where it 

 contributed to the atmospheric humidity, and that sodium chloride 

 actually protected the iron from rust by absorbing water vapour from 

 the air and thus reducing still further its humidity. The reason why 

 the sea-water salts did not produce the same effect as the common salt 

 solution was because of its extraordinary hygroscopic properties 

 maintaining an atmosphere with more moisture than would be the case 

 with a sodium chloride solution. It should be mentioned that during 

 the latter part of the experiment the inside of jar No. 2 was covered 



