
ag 
ON RAILWAY BAR CORROSION. 97 
London, was obtained prior to Experiment No. 1, and was now again used 
to adjust fourteen cast-iron 56lbs. weights by, so as to enable 7 ewt. of rails 
to be weighed at once. These cast-iron weights, prior to final adjustment, 
were coated with a thin covering of copal varnish, to preserve them from 
corrosion, until again called into use, after the lapse of two years. They 
were handled with leather slings to avoid abrasion, and preserved in a per- 
fectly dry place, and checked against the brass standard again before being 
used to weigh the rails after their removal. A steel beam of 36 inches 
& 
eo 
“a 
< 
in length was prepared and carefully adjusted by Mr. Yeates, instrument- 
maker, Dublin, by which these weights were adjusted. It was sensible to 
20 grs., with 56lbs. avoir. in each pan, or to zg} qth part of the whole 
load. It turned in all its bearings on hardened steel edges, and was found a 
most satisfactory and accurate instrument. 
A large beam was also prepared and carefully adjusted, whose length was 
eight feet, and whose strength was such as to be capable of weighing three 
lengths of rails at once, or of sustaining a load of above 12 ewt. This 
beam was of wrought iron, turning on hardened steel knife-edges, and with 
means of gradually bringing the load upon the beam without jar or vibration. 
When loaded with three rails and their counterpoise, this very large beam 
~-was found to be sensible to 500 grs., or to gggod part of the gross load; 
and could have been made still more so if requisite. It is probably the 
largest and most accurate beam ever made. . 
- Both this and the smaller beam were tolerably equibrachial ; but to avoid 
any error from this source, double weighings were made in adjusting the 
weights, and one end of the large beam was marked, and the rails always 
placed under it, the counterpoise being at the opposite end. 
This may appear a tedious description of an unnecessary amount of care ; 
but when it is recollected that the question to be determined relates to a 
weight of not much more than a single pound in a gross load of nearly 
1400lbs., it will be seen that any inaccuracy in the weighings would mate- 
rially modify, or wholly vitiate the results; and it is to the accumulation of 
slight errors of this sort, and probably more particularly to want of equibra- 
chiality in the beams used, or want of attention to always weighing at the 
one end, that I attribute the want of consistency of the results obtained on 
the Ulster line of railway. 
In order also further to increase the accuracy of the result, I proposed to 
allow a longer period to elapse before again removing the rails from the line 
when laid down. 
The same set of eighteen rails divided into three classes of six each, which 
had been used in the first experiment on the Kingstown line, were now again 
brought into requisition. They had lain since the former experiment hori- 
zontally under cover in a dry place, and had acquired a very slight coat of 
red rust. They were all placed in a boiler-maker’s oven, and exposed to a 
bright red heat, and then permitted to cool, without being exposed to any 
blows or jars, in a horizontal position and under cover. They had all now an 
uniform coat of black oxide (Hammerschlag), very thin and adherent, were 
pretty free from magnetism, except that due to terrestrial induction; and in 
this state they were all weighed, and the weights registered, each rail having 
Teceived a permanent mark at one end. 
_ The six rails to be exposed to abrasion only, were now heated horizontally 
to about 400° Fahr. and coated with boiled coal-tar, which rapidly dried into 
a tough japan varnish. 
The weighings were made on the 10th October 1842; and on the 18th 
oe 1842 they were placed upon the up line of the Dublin and Kingstown 
— 1849. H 
