-PROCEEDINGS OF SECTION J. Le 
a proper system of specifications and tests becomes a matter of 
the very highest practical moment. On the one hand, it is most 
important that bad material should be rigorously excluded, while 
on the other it is very desirable that the contractor or manu- 
facturer should not be harassed by unreasonable and impracticable 
requirements. 
Anyone unacquainted with the details of engineering practice 
might naturally suppose that this question had been threshed out 
and standard specifications universally adopted long since. Such, 
however, is not the case. The most inconsistent and various 
tests have been, and are still in not a few cases, employed—some, 
for example, so exacting that the very best and costliest qualities 
can hardly reach the standard ; and others so lenient that none 
but the most outrageously bad will be rejected. 
In the hope of arousing discussion on this very important 
question, and in some small measure helping to establish and 
promote the general adoption of tests that shall be at once con- 
sistent in themselves, reliable as securing the public safety, and 
not unduly oppressive in their effect on the manufacturer, I have 
in the present paper collected and commented upon such specifi- 
cations as have come under my own notice, and shall conclude by 
-proposing rules for general adoption. 
First, as to cast-iron. This material is usually tested as a 
beam, the test-piece being made either 1 inch square, or 2 inches 
deep and | inch wide, and slightly over 3 feet long, supported on 
knife edges 3 feet apart, and loaded at the centre. Such a test- 
piece is easily made, and the load required for fracture is so 
moderate that, in the absence of a proper testing-machine, extem- 
porised appliances of a simple and inexpensive nature will suffice 
—a practical convenience, as tests are often neglected which need 
elaborate apparatus. 
Now let us enquire what would be a reasonable central load for 
a beam | inch square and 3 feet span to carry. Rankine, in his 
“Qivil Engineering,” gives data from which I compute it at from 
470 to 1036 lbs. for different qualities, the former result being hot 
blast-iron that had been melted no less than eighteen times, while 
the latter was for the same at its twelfth melting. Good cold 
blast-iron gave 871 lbs. 
Trautwine gives from 500 to 900 lbs. as the strength of a 
similar beam, 700 being a fair average. 
Stoney, in his work on “Strains,” gives 847 lbs.; and a 
standard specification adopted by thirteen of the leading bridge 
manufacturers of the United States takes 500 lbs. as the test load 
of a bar 4 feet 6 inches span, which is equivalent to 750 lbs. on 
3 feet. 
The South Australian Railway Department, in several specifi- 
cations of comparatively recent date, makes 7 cwt., or 784 lbs., 
the standard ; while the Victorian Water Supply Department, in 
