all known Substmices to the Magnetic Injluencc, 4*c. 327 



feet within the front edge must be thicker by one inch than at 

 the edge. This correction, then, being applied to the distance 

 before measured, gives 3 feet 5^ inches as the real distance, 

 being only Jth of an inch different from the result of the compa^ 

 rative experiment — an error of only 5 J 5th part of the whole. 



II. This, with many other analogous results, being so sa- 

 tisfactory, IT NEXT BECAME AN OBJECT OF IMPORTANCE TO 

 ASCERTAIN— THE DISTANCE TO WHICH THE MAGNETIC IN- 

 FLUENCE MAY BE EMPLOYED FOR THE MEASUREMENT OF THE 

 THICKNESS OP SOLID SUBSTANCES. 



For the satisfactory determination of this question, however, 

 a number of distinct and particular investigations become ne- 

 cessary ; such as the relation which may exist between the in- 

 fluence exerted by different magnets, and their quality, dimen- 

 sions, and number; with the law of the directive power of 

 straight bar-magnets at different distances. 



1. In regard to the effect which tlce qiudity, dimensions^ and 

 number of magnets have on the extent of the infliience and the 

 accuracy of the results, a considerable series of experiments have 

 been undertaken. 



(1.) As to the quality of tlie magnets, I can only give the 

 general results which my imperfect experiments on this point 

 seemed to afford. Magnets constructed of steel made out of 

 the best Swedish iron, appeared to me, in each case where I 

 could compare them, to be susceptible of the greatest degree of 

 power. And those which were tempered throughout their 

 length (say reduced to a gold colour at the ends and gradually 

 softer, or down to a blue in the middle), were more powerful 

 than others, which were only hardened at the poles. 



The latter result may be considered as fully established, being 

 in accordance with all previous experiments by Coulomb, Biot, 

 Kater, and other magneticians ; but the former, as to the effect 

 of the quality of the steel, I state hesitatingly, because of its 

 being at variance with the conclusions of different careful expe- 

 rimenters. 



