24)6 MR. CHARLES o'nEILL ON CHANGES OF DENSITY 



Mean density of rolled copper S'SgS 



Mean density when-hammered 8 "87 8 



Mean density when annealed 8-896 



The other five were lost or not weighed. 



I made an experiment to try if this loss of density was 

 progressive, and depended upon the amount of hammering. 

 I took a piece of square bar of copper, 1*95 inches long 

 and 0'5 inch square, and hammered it myself upon an 

 anvil, taking its density in all eight times, each successive 

 hammering being more severe than the previous one. I 

 tabulate them as follows : 



Sp. gr. of bar copper (weight 1097 "35 grs.) ... 



Hammered slightly 



2McZ Hammering, length 2'io inch 



■^rd Hammering, length 2*20 inch 



ifth Hammering, length 2*50 inch 



^th Hammei'ing, length 3*00 inch 



dtli Hammering, length y^o inch 



']th Hammering, length 4*50 inch 



i>th Hammering, length 5-50 inch 



876 

 871 

 869 

 867 



It lost five grains in weight during the successive ham- 

 merings, and was several times made very hot. The last 

 hammering cracked the edges, and the recorded density 

 may not be correct on account of the difficulty of taking 

 it. The difference in density was O'OiS. 



In the two first series of experiments, by omitting the 

 third decimal when under five and adding one to the 

 second when above five, we get results perfectly alike in 

 each determination of the second series upon which I place 

 the most confidence;* and the effect of hammering and 

 annealing may be stated as follows : . 



^nd Series. ist Series. 



Eolled copper has a density of 8-90 8-88 



When hammered 8-88 8-86 



Hammered and annealed 8-90 8-88 



The decrease of density, then, is about i in 450; that 

 is, the copper has increased in bulk that much. The 

 cubical expansion of copper by heat is about 35W& foi" a 



* But in fact the weighings justify the three places of decimals. 



