428 Prof, dimming on the Development of [June, 



A brass bar with platina wire gave 10° negative deviation ; 

 with silver wire 20° positive ; with brass wire 15°, and with zinc 

 wire 25°, both positive. 



An iron bar with brazed copper wires, and the heat of two 

 lamps, gave 45° of negative deviation ; it likewise deviated nega- 

 tively with wires of platina. 



Two connecting wires, each composed of a platina and a 

 silver wire soldered together at one of their extremities, were 

 rivetted into a copper bar : when the silver ends were rivetted 

 to the bar, the deviation was positive ; when the platina, nega- 

 tive; but when the platina ends were shortened to half an inch, 

 the deviation again became positive. 



Quicksilver inclosed in a glass tube, eight inches long, and 

 half an inch in diameter, with copper wires, gave at the temper- 

 atures 170° and 115° the corresponding deviations 8° and 3° 

 with the small compass. 



A double bar, eight inches long, composed by soldering 

 together at the middle a bar of bismuth and another of anti- 

 mony, when heated at both extremities, and kept at a tempera- 

 ture of 60° in the middle, gave a deviation of 36° with the large 

 compass at the melting point of bismuth. 



An alloy of bismuth and antimony gave with the large com- 

 pass a deviation of 3° negative; on continuing the heat it 

 returned to zero, and at the melting point of the bar, it became 

 4° positive. 



A bar of bismuth, six inches long, was broken into two 



a b c d 



parts ; I | | , a b and c d of two 



inches and four inches in length, and soldered together again 

 with a thin plate of copper interposed. When heated at a or c, 

 the deviation was positive i and when at b or d, was in the oppo- 

 site direction; but when the bar was unbroken, the deviations 

 had been in the same direction, whether the heat were applied 

 at a, b, or c. 



A rod of bismuth having wires soldered to it in different 

 parts of its length, the included portions being alternately hot 

 and cold, exhibited as many poles as there were wires. 



Two bars, whose deviations were respectively 20° and 16°, 

 when connected in sequence gave a deviation of 23° ; the devia- 

 tion was not so great when the hot and cold ends were con« 

 nected respectively. 



A battery of eight bars produced some, but not a consider- 

 able increase of power. 



A bar which produced a deviation of 20° with four feet of 

 copper wire, of 1-21 of an inch diameter, gave with 8, 16, and 32 

 feet of the same wire, corresponding deviations of 15i-°, 10°, 

 and 7°. 



With eight feet of copper wire, of l-37th inch, the deviation 



