Molecular Change in Iron Wire. 63 



to a needle with index-plate, as in the figure, conspicuous effects 

 were obtained ; but the momentary elongation was relatively much 

 less (in one instance ^j of the length of the heated part) than 

 when a battery was employed, apparently in consequence of the wire 

 being less intensely heated. 



A large number of experiments were made with wires of palla- 

 dium, platinum, gold, silver, copper, lead, tin, cadmium, zinc, brass, 

 german- silver, aluminium, and magnesium (wire and ribbon), dimi- 

 nishing the length and thickness of the wire in each case, and ad- 

 justing the tension until suitable temperature and strain were obtained; 

 but in no instance could a similar molecular change to that observed 

 in iron be detected. Palladium and platinum wires of different 

 lengths, thickness, and degrees of strain were examined at various 

 temperatures, up to that of a white heat ; but no irregularity of co- 

 hesion, except that of gradual softening at the higher temperatures, 

 was observed ; they instantly contracted with regular action on stop- 

 ping the current. Several gold wires were similarly examined at dif- 

 ferent temperatures up to that of a full red heat ; no irregularity 

 occurred either during heating or cooling ; but little tension (about 

 4 ounces) was applied, on account of the weak cohesion of this metal. 

 Wires of silver similarly examined would only bear a strain of about 

 2 ounces, and a temperature of feeble red heat visible in daylight ; 

 no irregularity of elongation or contraction occurred during heating 

 and cooling. By employing exactly the proper temperature and 

 strain, a very interesting phenomenon was observed : the wire melted 

 distinctly on its surface without fusing in its interior, although the 

 surface was most exposed to the cooling influence of the air ; this oc- 

 curred without the wire breaking, as it would have done if its interior 

 portion had melted : the phenomenon indicates the passage of the 

 electricity by the surface of the wire in preference to passing by its 

 interior. Wires of copper expanded regularly until they became red- 

 hot ; they then contracted slightly (notwithstanding the strain ap- 

 plied to them), probably in consequence of a cooling effect of in- 

 creased radiation produced by the oxidized surface, as a similar effect 

 occurred with brass and germ an- silver*. On stopping the current 

 the wire contracted without manifest irregularity. Wires of lead and 

 tin were difficult to examine by this method, on account of their ex- 

 tremely feeble*eohesion and the low temperature at which they soft- 

 ened : wires about 1"63 millimetre diameter, 25*5 centimetres long 

 (with a strain upon them of about one ounce), were employed ; no 

 irregularity was detected. Wires of cadmium from 1*255 millimetre 

 to 1*525 millimetre thick, and 24*2 centimetres long (with a strain 

 of two ounces), exhibited a slight irregularity of expansion at the 

 lower temperatures ; they elongated, and also cooled, with extreme 

 slowness, more slowly than those of any other metal. Wires of zinc 

 exhibited a slight irregularity of expansion, like those of cadmium ; 

 the most suitable ones were about 25 centimetres long and 1*2 mil- 

 limetre in diameter, with a strain of 10 ounces. Wires of brass and 



* This supposition does not agree with the results obtained with iron wire, 

 which also oxidizes freelv. 



