718 PROCEEDINGS OF SECTION J. 
their specification for the Loddon Weir, are satisfied with 6 ewt., 
or 672 lbs. Some years ago I had the opportunity of testing a 
large number of such bars of iron from four different Melbourne 
foundries, using different brands of pig, and in one case old 
American railway wheels. More than 60 bars were broken, half 
being sand-castings and half chilled, and the results were very 
uniform. In no case did a sound bar carry less than 700 lbs., 
while in only five cases did a sand-casting carry more than 900, 
and in only two more than 1000 lbs., and but one chilled bar 
reached 900 lbs. These experiments correspond well with the 
South Australian specification, only six bars out of thirty-one 
falling below it, and five of these being only a few pounds 
deficient. If the test had been reduced from 7 ewt. to 750 lbs., 
the American figure, only one bar that was not visibly defective 
would have failed to pass, and this test I would suggest as a 
reasonable one to adopt in practice. 
Some engineers prefer a larger test bar, and specify one 2 
inches deep by 1 inch wide. J am inclined, on the whole, to 
join in this preference, on the ground that such a bar approaches 
more nearly to the dimensions of the objects for which the iron 
is intended. The Victorian Railway Department for many years 
has employed this size, and fixes 30 cwt. as the load, and so does 
the Victorian Water Supply in their Goulburn Weir specification, 
while the New South Wales Railway Department in the case of 
Penrith Bridge asked for 27 cwt.. Now, all my experience leads 
me to regard a 30 ewt. test as impracticably high. Theory would 
lead us to conclude that bars of this latter size would be four 
times as strong as the inch bars previously referred to, in which 
case 28 cwt. or 3136 lbs. would be a fair test, corresponding to 
that used’ in South Australia. All my experience with colonial 
castings, however, leads me to the conclusion that a 2-inch by 
l-inch bar is far from being four times as strong as one an inch 
square, a conclusion that is not very astonishing when one con- 
siders the nature of the material and the hardening effect of the 
rapid cooling of the external skin. The results I have so far 
obtained range for sound sand-castings from 2116 to 2850 Ilbs., 
and I do not think the standard should be fixed higher than 
2450 lbs., or say 22 cwt., which is barely 3°3 times as much as 
the smaller bars endure. Our president, Professor Warren, 
informs me that he has never known such a bar to carry more 
than 28 ewt., or 3136 lbs., and that 25 cwt. is a good test. My 
own experience, as above stated, leads me to a still lower figure. 
Some engineers specify not only a test load, but also a 
deflection that must be reached before fracture, arguing that a 
very hard iron might carry a large weight gradually applied, but 
at the same time be so deficient in resilience as to be very easily 
fractured by a blow. This view is, I think, reasonable, and from 
my own experience I would suggest a 23-inch as the least allow- 
