408 



PROCEEDINaS OF THE AMERICAN ACADEMY 



observed that pillars with rounded ends generally broke at the middle 

 only; pillars with flat ends generally broke in three places, at the mid- 

 dle and near each end. 



In 1856, Hodgkinson presented a second paper, on the strength of 

 cast-iron pillars, to the Royal Society.* Its special subject is the 

 strength of pillars cast from iron from various parts of Great Britain. 

 The experiments were made at University College, London, with a 

 larger and more powerful apparatus than that which he previously 

 used at Manchester. The pillars tested were longer and larger, and 

 the results reported afford the means of testing, in a very satisfactory 

 manner, the formulas previously determined. The earlier experiments 

 were undertaken for the purpose of determining the laws according to 

 which the breaking weights of pillars depended upon their dimensions. 

 For this purpose, it was necessary to adhere to one description of iron. 

 A Yorkshire iron, Low Moor, No. 3, was selected, which is described 

 as a good iron, not very hard. The formulas were satisfactorily de- 

 tei'mined fcjr this iron, but there was some risk in applying them to 

 l^illars cast from otlier descriptions of iron, without lurther experi- 

 ments. In the second paper, experiments are described on pillars cast 

 from thirteen kinds of iron, — two of them being mixtures. 



For the purpose of effecting a systematic comparison between the 

 breaking weights by the experiments of 185G and the breaking weights 

 computed by the formulas of 1840, the three following tables have been 

 computed, the data being extracted from Hodgkinson's paper of 1856. 



Table I. 



* Experimental Researches on the Strength of Pillars of Cast-iron, from various 

 Parts of the Kingdom. By Eaton Hodgkinson. Philosophical Transactions of the 

 Royal Society of London for 1857. Part III. 



