lo4 Liout.-Ool. A. R. Clarke on the Elasticity of Brass. 



The probable errors hero refer to the last figure of the pre- 

 ceding decimal. An examination of this Table shows that 

 there are, notwithstanding the extreme care with which the 

 observations were made, some unknown constant errors in ex- 

 istence outweighing the apparent probable errors of the differ- 

 ent measurements, and therefore interfering with the applica- 

 tion of the theory given above, at least for rods Nos. 1 and 2. 

 I hope by some future observations to discover the source of 

 the constant errors in question. The following Table gives the 

 values of E for each rod expressed in millions of pounds. 



Rod. 



Face I. 



Face III. 



Face II. 



Face IV. 



Mean. 



No. 1. 

 „ 2. 

 „ 3. 

 „ 4. 



16-806 

 16-945 

 16181 

 15 083 



16-769 

 16992 

 16-181 

 15105 



16-737 

 16-937 

 16192 

 15076 



16-726 

 16960 

 16196 

 15 045 



16-759 

 16-958 

 16188 

 15077 



While making these experiments I met with a very elaborate 

 work published at St. Petersburg, entitled Recherches Expe'ri- 

 mentales sur V elasticite des Metaux faites a V observatoire phy- 

 sique central de Russie, par A. T. Kupffer. Of brass, nine spe- 

 cimens were experimented on — three of cast brass, three of 

 plate brass, and three of hammered. The experiments were 

 made : — first, by causing the rods to bend under given weights, 

 and measuring with vertical circles the inclinations of the ends 

 of the rods so bent ; secondly, by causing the rods to vibrate 

 transversely when fixed by one end in a vertical position, the 

 other end being weighted, and counting the vibrations. The 

 results obtained by the two methods agreed very satisfactorily. 

 The following Table gives the moduli of elasticity expressed in 

 millions of pounds for the nine specimens. 



Cast. 



Rolled. 



Hammered. 



12 262 

 11-537 

 14540 



15-338 

 16 244 



15-884 



16012 

 16523 

 15753 



These quantities result from a very large number of obser 

 vations. 



