•pd 35, exce|jt 26; pieces of each fort were heated both in 

 the flame of a candle, and in fand; all of which retained 

 the facalty under confideration after being treated in both 

 ways. In reality, wires that had been thus foftened, feemed 

 to be in the fame condition with fmall nails of caft iron, con- 

 iidered as retainers of magnetifra, though the latter are of a 

 much harder quality ; for a nail of the fort called fparrow- 

 bills by (hoe-makers CKhibited the appearances defcribed in 

 the fccond experiment, after being filed down to the thicknefs 

 of a fraall wire. 



If then that kind of magnetifm which I have ventured to 

 call conditional do not depend on comparative hardnefs, to 

 what caufe is the phenomenon to be defcribed? little can be 

 offered on my part, befides probable conjecture, in anfwer 

 to this quellion. The temperature of wire is confiderably 

 raifed during its paflage through the wordle ; and may not we 

 inwgine with fome ftiew of reafon, that this encreafe of tem- 

 perj^tiire, joined to the fubfequent contadl of cold air, pro- 

 duces ^ new arrangement of the molecules conflituting the 

 wire which enables it to retain a portion of magnetifm as long 

 as it remains in conta6l with a ferruginous body ? if this 

 fuppofition be true, experiment proves the new "arrangement 

 to take place in the 22 wordle ; when the flendernefs of the 

 wire will occafion it to cool fuddenly after paffing through the 

 inftrument. The reality of fuch changes in the texture of 

 bodies which are not in a ftate of fufion, is admitted at 

 prefent by experimental philofophers. I may alfo quote in 

 favour of this hypolhetis fome valuable obfervations made by 

 Gregory Watt, Efq, on the various degrees of magnetifm 

 exhibited by the fame bazaltic ftone under different forms of 

 eryftallization ; which obfervations may be feen in your 

 Journal for February, 1805. 



Any attempt to explain the permanent magnetifm of fmall 

 wires during their connedion with foft iron, and the lofs of 

 this properly which enlues when the connection is broken, 

 appears to be fuperfluous, becaufe the fadl is evidently / 



analogous to the well known method of adding flrength to a 

 magnet by a gradual encreafe of its load j for this operation, 

 when judicioufly conduded, gives a magnetic charge to a 

 bar of fteel already touched, which it cannot retain after the 

 ^veight is removed. "'; . 



I remain, &c. ^'Ji 



JOHN GOUGH. 

 P. S. I neg. 



