182 Dr. Tyndall on Diamagnetism and Magneerystallic Action, 



be so. Bismuth powder will remain crystalline, and carbonate 

 of lime is never free from suspicion. I thought 1 had found an 

 unexceptionable substance in chalk, inasmuch as Ehrenberg has 

 proved it to be a mere collection of microscopic shells ; but Prof. 

 Ehrenberg himself informs me, that even these shells, which 

 require a high magnifying power to render them visible, are in 

 their turn composed of infinitesimal crystals of calcareous spar* 

 In this dilemma one way remains open to us : we will allow the 

 objection to stand, and follow it out to its inevitable conse- 

 quences ; if these are opposed to fact, the objection necessarily 

 falls. 



Let us suppose the bismuth powder to be reaiTanged, so that 

 the perfect crystal from which it was obtained is restored. In 

 this case the axes of all the little component ci*ystals are parallel^ 

 they work all together, and hence their action must be greats 

 than if only a majority of them were parallel. In a bismuth 

 crystal, therefore, the difference of action in the line of the mag- 

 necrystallic axis, and in a line perpendicular thereto, must be a 

 maximum. It must, for example, be greater than any difference 

 which the model of bismuth powder can exhibit ; for a portion 

 of the force attributed to the axes must in this case be annulled 

 by the confused grouping of the little component ciystals. In 

 the words of Professor Thomson, it is merely a balance of action 

 brought about by predominance, which can make itself manifest 

 here. Hence if we measure the repulsion of the crystal in a 

 direction parallel to the principal cleavage, and in a direction 

 perpendicular to it, and also measure the repulsion of the model 

 in the line of compression and in a line perpendicular to it, the 

 ratio of the two former repulsions, that is, of the first to the 

 second, must be greater than the ratio of the two latter, that is, 

 of the third to the fourth. 



Turning to Tables IX. and X., we see that the ratio of the 

 repulsion of the crystal in the direction of principal cleavage to 

 the repulsion in a direction perpendicular to the same is ex- 



15 



pressed by the fraction yy =1-36. Turning to Tables XL and 



XII., we find that the ratio of the repulsion of the model in the 

 line of compression to the repulsion in a line perpendicular to it 



3 



is expressed by the fraction ^ =1*5. In the latter case, there- 

 fore, we have the greatest differential effect ; which result, were 

 the repulsion due to the mere predominance of axes, as urged by 

 Mr. Thomson, would certainly bear a suspicious resemblance to 

 the conclusion that a part is greater than the whole. This result 

 has been entirely unsought. The models were constructed with 



