1869.] 



Molecular Change in Iron Wire. 



263 



loped ; and if too great, the wire was overstretched or broken when it 

 approached the softening-point. The amount of strain imparted was ap- 

 proximately measured by temporarily substituting the body of a small 

 spring balance for the hooked wire F. The heated wire must be protected 

 from currents of cold air. 



With wires of iron 0'65 millimetre thick (size " No. 23") and 21*5 

 centimetres long, strained to the extent of 10 ounces or more, and heated 

 to full redness, the phenomenon was clearly developed. As an example, the 

 needle of the instrument went with regularity to 18*5 of index-plate; the 

 current was then stopped ; the needle instantly retreated to 1 7*75, then as 

 quickly advanced to 19*75, and then went slowly and regularly back, but 

 not to zero. If the temperature of the wire was not sufficiently high, or the 

 strain upon the wire not enough, the needle went directly back without ex- 

 hibiting the momentary forward movement. The temperature and strain 

 required to be sufficient to actually stretch the wire somewhat at the higher 

 temperature. A higher temperature with a less degree of strain, or a 

 greater degree of strain with a somewhat lower temperature, did not deve- 

 lope the phenomenon. The wire was found to be permanently elongated 

 on cooling. The amount of elongation of the wire during the momentary 

 molecular change was usually about part of the length of the heated 

 part of the wire ; but it varied in different experiments ; it was greatest in 

 amount when the maximum degrees of strain were applied. The mole- 

 cular change evidently includes a diminution of cohesion at a particular 

 temperature during the process of cooling ; and it is interesting to notice 

 that at the same temperature during the heating -process no such loss of 

 cohesion (nor any increase of cohesion) takes place ; a certain temperature 

 and strain are therefore not alone sufficient to produce it ; the condition 

 of cooling must also be included. The phenomena which occur during 

 cooling are not the exact converse of those which take place during heating. 



The phenomenon of elongation of iron wire during the process of cool- 

 ing evidently lies within very narrow limits ; it could only be obtained 

 (with the particular battery employed) with wires about 21*5 centimetres 

 (=8-^ inch) long, and about O'Go millimetre ( = Nos. 22 & 23 of or- 

 dinary wire-gauge) thick, having a strain upon them of 10 ounces or 

 upwards ; with a weaker battery the phenomenon could only be obtained 

 by employing a shorter and thinner wire. 



The experiment may easily be verified in a simpler manner by stretching 

 an iron wire about TO millimetre diameter between two fixed supports, 

 keeping it in a sufficient and proper degree of tension by means of an 

 elastic band^ then heating it to full redness by means of a row of Bunsen's 

 burners, and, as soon as it has stretched somewhat, suddenly cutting off the 

 source of heat. In some experiments of this kind, with a row (42 centi- 

 metres long) of 21 burners and a row (76 centimetres long) of 43 burners, 

 and the wire attached to a needle with index-plate, as in the figure, con- 

 spicuous effects were obtained ; but the momentary elongation was relatively 



