174 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



3. If these surfaces diminish, the tension augments up to 

 the point when they become insufficient to allow the element- 

 ary poles to expand, and a portion of two contrary magnet- 

 isms disappears in order to reproduce the natural state. 



4. If the surfaces change, but the mean sections remain 

 constant, the quantity of magnetism remains constant, al- 

 though its distribution may have varied very considerably 

 with the change in the surface. If the surface remain con- 

 stant, but the sections vary, the quantities of magnetism are 

 proportional thereto. If the surface is not great enough, the 

 total magnetism, which increases with the area of the section 

 of the magnet, will be unable to expand itself over the bar, 

 and must diminish itself; thus establishing the limit beyond 

 which it is not possible to magnetize a given bar of iron. 



On the other hand, if the mean section of the magnet di- 

 minish indefinitely, the tension of the magnetism will not in- 

 crease indefinitely, but will cease to increase at a given limit 

 fixed by the relation between the mean section and the sur- 

 face of the bar. Having established the fact that the mean 

 section determines the quantity of the magnetism, and that 

 the surface of the masjnet is that which regulates the distri- 

 bution, it becomes evident that by a proper arrangement of 

 the surface the latter may be saturated ; that it is, in fact, 

 the saturation of the surface which determines the saturation 

 of the magnet. If, then, the magnet is made greater without 

 increasing the quantity of magnetism, the tension of the lat- 

 ter becomes feebler; if it is made smaller, a portion of the 

 magnetism is lost, but if, in this case, Ave add to the poles a 

 piece of unmagnetized iron, we afford the space needed for 

 the retention of the magnetism, and thus retain the strength 

 of the magnet: this latter conclusion is well verified by di- 

 rect experiment, and explains a portion of the action of arma- 

 tures. G C\ LXXVIII., 1500. 



DIELECTRIC ABSORPTION. 



Faraday, as is well known, discovered that a condenser 

 takes on a decidedly greater quantity of electricity when one 

 of the plates is charged with electricity while the other is 

 connected with the earth. The quotient of the quantity of 

 electricity which the isolated plate assumes when the sep- 

 arating non-conductor is air, and when the non-conductor is 



