II .
" Experimental Researches on the Strength of Pillars of Cast Iron .
" By EATON HODGKINSON , Esq. , F.R.S. , Professor of the Mechanical Principles of Engineering , University College , London .
Received November 20 , 1856 .
( Abstract .
) In a previous paper oil this subject ( Philosophical Transactions , 1840 ) , I had shown , -ist , that a long circular pillar , with its ends flat , was about three times as strong as a pillar of the same length and diameter with its ends rounded in such a manner that the pressure would pass through the axis , the ends being made to turn easily , but not so small as to be crushed by the weight ; 2nd , that if a pillar of the same length and diameter as the preceding had one end rounded and one flat , the strength would be twice as great as that of one with both ends rounded ; 3rd , if , therefore , three pillars be taken , differing only in the form of their ends , the first having both ends rounded , the second one end rounded and one flat , and the third both ends flat , the strength of these pillars will be as 1-2-3 nearly .
The preceding properties having been arrived at experimentally , are here attempted to be demonstrated , at least approximately .
The pillars referred to in my former paper were cast from Low Moor iron No. 3 ; they were very numerous , but usually much smaller than those used in the present trials .
I felt desirous too of using the Low Moor iron in the hollow pillars employed on this occasion , not on account of its superior strength , but its other good qualities .
The pillars from this iron were cast 10 feet long , and from 2to 4 inches diameter , approaching in some degree , as to size , to the smaller ones used in practice .
The results from the breaking weights of these were moderately consistent with the formulae in the former paper , with a slight alteration of the constants , rendered necessary by the castings being of a larger size , and therefore softer than before , a matter which will be adverted to further on .
The formulae for the strength of a hollow pillar of Low Moor iron No. 2 , -where w is the breaking weight , in tons , of a pillar whose length is I in feet , and the external and internal diameters D and d in inches , the ends being flat and well bedded-are as below : D 3'55d3'55 w-46 65 xfrom formula in Phil. Trans. 1840 ; D3-5 -d3-5 w=42347 x 1163 from formula in present paper .
To obtain some idea of the relative strengths of pillars of different British irons , I applied , at Mr. Stephenson 's suggestion , to Messrs. Easton and Amos , who procured for me twenty-two solid pillars , each 10 feet long and 2inches diameter , cast out of eleven kinds of iron ( nine simple irons and two mixtures ) .
The pillars were all from the same model , and were cast vertically in dry sand , and turned flat at the ends , as the hollow ones had been ; two being cast from the same kind of iron in each case .
The simple unmixed irons tried were as below , and all of No. 1 .
Mean breaking weight .
Old Park iron ... ... ..Stourbridge ... ... 29'50 tons .
Derwent iron ... ... ..Durham ... ... ... .
28*03 , , Portland iron ... ... .
Tovine , Scotland ... 27'30 , , Calder iron ... ... ... Lanarkshire ... ... . .
27'09 , Level iron ... ... ... .
Staffordshire ... ... 24 6 7 , , Coltness iron ... ... Edinburgh ... ... . .
2352 , , Carron iron ... ... ... Stirlingshire ... ... .
2352 , , Blaenavon iron ... . .
South Wales ... ... 2205 , , Old Hill iron. .
S. .
Staffordshire ... . .
20'05 , , The mean strength of the pillars from the irons above varies from 20-05 to 29-50 tons ; or as 2 to 3 nearly .
The pillars formed of mixed irons were found to be weaker than the three strongest of the unmixed series .
From many experiments , it was shown that the weight which would crush the pillars , if they were very short , would vary as 5 to 9 nearly .
The pillars in general were broken of four different lengths , 10 feet , 7 feet 6 inches , 6 feet 3 inches , and 5 feet , the ends of all being turned flat , and perpendicular to the axis .
It was found that when the length was the same , the strength varied as the 3*5 power of the diameter ; and when the diameter was the same and the length varied , the strength was inversely as the 1'63 power of the length .
Both of these were obtained from the mean results of many experiments .
The formula for the strength of a solid pillar would therefore be d3m5 w=m x l---3 where w is the breaking weight , d the diameter in inches , Ithe length in feet , and ma weight which varied from 49'94 tons in the strongest iron we tried , to 33'60 tons in the weakest .
The ultimate decrement of length , in pillars of various lengths but of the same diameter , varies inversely as the length nearly .
Thus the ultimate decrements of pillars 10 feet , 7 feet 6 inches , 6 feet 3 inches , and 5 feet , vary as 2 , 3 , 3 } and 4 nearly , according to the experiments , from which it appeared that the mean decrement of a 10-feet pillar was ?
176 inch .
Irregularity in Cast Iron .
The formulae arrived at in this paper are on the supposition that the iron of which the pillars are composed is uniform throughout the whole section in every part ; but this was not strictly the case in any of the solid pillars experimented upon .
They were always found to be softer in the centre than in other parts .
To ascertain the difference of strength in the sections of the pillars used , small cylinders ; inch diameter and 1inch high , were cut from the centre , and from the part between the centre and the circumference , and there was always found to be a difference in the crushing strength of the metal from the two parts , amounting perhaps to about one-sixth .
The thin rings of hollow cylinders resisted in a much higher degree than the iron from solid cylinders .
As an example , the central part of a solid cylinder of Low Moor iron No. 2 , was crushed with 29'65 tons per square inch , and the part nearer to the circumference required 34'59 tons per square inch ; cylinders out of a thin shell half an inch thick , of the same iron , required 39'06 tons per square inch ; and other cylinders from still thinner shells of the same metal , required 50 tons per square inch , or upwards , to crush them .
As these variations in cast iron have been little inquired into , except by myself , and have never , so far as I know , been subjected to computation , I have bestowed considerable trouble upon the matter , in an experimental point of view , and endeavoured to introduce into the formulae previously given , changes which will in some degree include the irregularities observed .

