110 REPORT—1849. 
Again, on the other hand, the existence of polarity of variable magnetic 
intensity in every rail involves the existence of electric currents circulating 
Fig. 11. 
———___“e ni 
———— 
@. * 
“ b> 
| 
round the bars, in accordance with the facts pointed out so well by Dr. Fa- 
raday. And in accordance with the recent experiments of Mr. Grove, the 
constantly recurrent induction of magnetism of great intensity in each rail 
involves a constant change of temperature in the rail, due to this cause alone, 
and probably an equally constant change in the molecular arrangement of 
its particles. 
Such are the facts, so far as I have been enabled to observe them, of the 
complex relations to electricity, magnetism, and terrestrial temperature of 
railway bars ; they fail to throw any direct light upon the immediate subject of 
our inquiry, but since the closest relationship has been proved to subsist be- 
tween all these molecular forces, and especially since the later refined re- 
searches of Faraday and Pliicker have shown that changes in the electrical 
or magnetic state of solids is attended with an instantaneous ehange in the 
relative groupings of their molecules, and knowing beforehand that chemical 
action in its most ordinary circumstances is powerfully influenced and modi- 
fied by the state of grouping, or of aggregation of these molecules, it seems 
by no means improbable that the chemical action of air and moisture upon 
the iron of railway bars may be more or less modified by the electrical and 
magnetic forces that specially apply to them. To reduce this te certainty, 
demands experiments conducted, not after the manner or with the immediate 
object of those before us, but by refined research in the physico-chemical 
laboratory. . 
Interesting as such researches may be to science, and to which the facts 
here recorded may perhaps serve as finger-posts at the commencement of the 
way, they are not of very high value to the practical railway engineer, inas- 
much as we have already found that the destruction of railway bars by cor- 
rosion is small in comparison with that by traffic. Nor are we obliged to 
rest in any vague speculation to find efficient causes sufficient to account 
for the real difference that we have established between the corrosion of the 
same railway bar in use and out of use. 
The three principal causes to which I attribute this difference are,— 
Ist. The top surface of a railway bar in use is constantly preserved in a 
state of perfect cleanness, polish, and freedom from oxidation, while the re- 
mainder of the bar is rough, coated originally with black oxide (Ham- 
merschlag) and soon after with red rust (peroxide and basic salts). 
Not only is every metal electro-positive to its own oxides, but, as esta- 
blished in my SecondReport on the Action of Air and Wateron Iron (sec. 242), 
a mass of metal, partially polished and partly rough, is primarily corroded on 
the rough portion. Hence then a railway bar while in use is constantly pre- 
served from rusting by the presence of its polished top-surface. Such polished 
surface has no existence upon the rail out of use. 
2nd. The upper portion of the rail in use is rapidly condensed and hard- 
ened by the rolling of traffic over it; and I have also shown in the same 
reports, that all other circumstances being the same, the rate of corrosion of 

