ON RAILWAY BAR CORROSION. 111 
any iron depends upon its density, and is less in proportion as this is ren- 
dered greater by mechanical means. 
8rd. As every metal is positive to its own oxides, the adherent coat of rust 
upon iron, while it remains, powerfully promotes the corrosion of the metal 
beneath, and this in a greater degree in proportion as the rust adherent is of 
greater antiquity, inasmuch as it has been shown that the rust produced by air 
and moisture, which at first contains but little peroxide (Fe, O, ), continues to 
change slowly, and becoming more and more peroxidized, becomes more and 
more electro-negative to its own base. 
Now the rust formed upon a railway bar out of use continues always to 
adhere to it, and thus to promote and accelerate its corrosion, while the rust 
formed upon a railway bar in use is perpetually shaken off by the vibration 
of traffic, and thus this source of increased chemical action is removed. Of 
the extent to which this acts, we are informed by the results of the second 
experiment on the Kingstown Railway, where the weight of adherent rust, 
formed on removal of the bars out of use, was found to be more than four 
times as great as that upon an equal surface of the bars that had been in use. 
We have found that this difference of corrosion in and out of use however 
is a constantly decreasing one; this arises from the fact that the condensation 
of the top-surface of the rails ceases after it has reached a certain point de- 
pendent on the maximum weight of the trains, and that after the lapse of a 
considerable period a uniform coat of rust is formed upon the rails in use, 
which is so firmly adherent that the vibration and wind of passing trains are 
incapable of sweeping it away; and it seems possible that after the lapse of 
an extremely prolonged time, the difference between the corrosion of the 
rail in use and out of use might become so small as to be perhaps insensible. 
To recapitulate, then, we have found that railway bars forming part of a 
Jong line, whether in or out of use, corrode less. for equal surfaces than a 
short piece of the same iron similarly exposed ; that the rails in use do cor- 
rode less than those out of use; that this difference is one decreasing with 
the lapse of time; that the absolute amount of corrosion is a source of de- 
struction of the rails greatly inferior to that due to traffic; that it is highly 
probable that the electrical and magnetic forces developed in the rails by 
terrestrial induction and by rolling traffic, react in some way upon the che- 
mical forces concerned in their corrosion ; and that therefore the direction of 
lines of railway in azimuth is not wholly indifferent as respects the question 
of durability of rails. 
I am not aware of any information upon this subject having any character 
of accuracy which I can refer to extraneous to the present report. In the 
Franklin Journal, and also in Silliman’s Journal, some few papers occur 
giving rather long statements as to the wear of rails on the Lowell and other 
American railways, as also some such in the Mining Journal (London), as 
to some English lines ; as also some observations upon the abrasion of cast- 
iron rails, given by Thompson in his ‘Colliery Inventions and Improve- 
ments. They need not however be extracted here. 
I might also extend this report to a comparison of the relations which 
subsist between the surface per yard lineal exposed to corrosion, and the 
strength and stiffness of the several principal forms of rail in use upon one 
line; but this can so obviously be done by the engineer for his own purpose, 
that it seems needless. It is however an element of choice in the form of 
rails that appears heretofore to have been wholly neglected. 
_ Iconclude therefore with two practical suggestions, deducible from the 
_ information we have obtained, having for object the increasing the durability 
_ of rails, both as against traffic and corrosion. First. Of whatever quality of 
