an Explanation of Hall's Phenomenon. 257 



that the transverse force would tend to make the metallic strip 

 assume a distorted form of the same character as that indicated 

 in fig. 3. As thus distorted, the strip may be mapped out 

 into six districts, in three of which the metal is, on the whole, 

 subjected to longitudinal traction, while in the other three it 

 undergoes compression. This, of course, supposes that the 

 metal, though cemented to glass, is capable of yielding to 

 some extent to the force acting upon it. And since neither 

 the metal itself, nor the glass, nor the cement which connects 

 them, is absolutely rigid, this must necessarily be the case, 

 though probably in a very small degree. The ends C, D of 

 the strip, which are in a weaker part of the magnetic field, or 

 even removed from its influence altogether, will not be acted 

 upon in the same manner ; and if the strip is sufficiently long 

 they will remain absolutely at right angles to the middle line, 

 as shown in the figure. Two points of contrary flexure will 

 therefore exist on each of the edges; and the necessary conse- 

 quence of this will be alternate regions of compression and 

 extension, as indicated in the figure. It seems to me that, 

 however thin the metal may be, so long as it is not altogether 

 without cohesion and elasticity, the action of a transverse 

 force across the middle will affect it in some such manner as 

 that pointed out, to some degree which may be very small. 



I have prepared a rough model to illustrate my idea of the 

 nature of the distortion which the plate undergoes. A num- 

 ber of narrow strips of deal are attached at short intervals to 

 a rectangular sheet of indiarubber: they are equal in length, 

 and parallel to the shorter sides of the rectangle. The two 

 strips at the ends of the sheet are nailed firmly to a board; 

 otherwise the indiarubber is free. If the middle strip of 

 wood is drawn or pressed in a direction transverse to the rect- 

 angle, the indiarubber is distorted and assumes the form of 

 fig. 3 ; and the regions which are there marked as stretched 

 and compressed appear very clearly, the strips being drawn 

 apart in the former, and crowded together in the latter. The 

 same kind of distortion would be produced, though to a 

 smaller extent, if the ends, instead of being rigidly fixed, were 

 able to yield, more or less, to pressure or tension. But unless 

 they were each attached to a single point, and were otherwise 

 perfectly free (in which case the sheet would be longitudinally 

 divided into two equal regions of compression and extension), 

 some such distribution of strain as that indicated in fig. 3 

 must necessarily occur*. 



* It should be remarked that no account whatever is taken in the 

 model of lateral compression and extension, which, as already stated, are 

 probably very small in the case of the metallic sheets used in Hall's 

 experiments. 



