Prof. Norton on Molecular Physics. 387 



of this induced electric polarization, and of the reflex action just 

 noticed, we may discern the probable origin of those supposed 

 variations of axial attraction which Professor Dana has shown 

 will suffice for the explanation of secondary planes in crystal- 

 lization. 



To illustrate by a special case, let 

 fig. 2 represent a process of crystalli- ^P 

 zation in which the particles are ar- ^/ 

 ranged in successive squares. When ^c 



c, c, c are the outer particles, their \ft r-^J~ (^*"igr~'|p 

 outer sides will be positively pola- ^g' J^ < « ^ie ^e* 

 rized, and will consequently exert an ^/~"^P"""^P"""^P"' ; 

 electric attraction, in addition to the ^. r j^ /i£ J& <g^ 

 increasing molecular attraction that *S?'~^F~' 

 results from the cooling, upon the Jy? jd£ r 

 molecules c', c', c' immediately exterior Wf~Wf~~ 

 to them. The next step in the pro- 

 cess should be the union of the molecules thus attracted ; and 

 all of these molecules should have the same tendency to unite, 

 unless there should be a material inequality in the physical con- 

 dition of the outer particles c } c, c, c on different sides of the 

 square, c c c c. But the four corner particles, c", c n , c", c' f , can- 

 not thus be directly brought into union with the particles of the 

 crystal. They must either unite with the nearest particles of 

 those newly attached, or remain disunited, to become incorpo- 

 rated at a later stage of the process. In the normal or complete 

 growth of the crystal, the first would be the result. If, on the 

 other hand, the forces of the new lines of outer particles c\ c' c J , c 1 

 should fall off materially in intensity, so that the corner particles 

 c 1 ' are not taken up in the same step of the process as the others, 

 secondary lines would arise at the angles of the square c 1 , c", c' } c". 

 In the whole cubical crystal, of which fig. 2 represents a section, 

 secondary planes would be formed at the edges. Every such line 

 or plane may have different positions, according as the corner par- 

 ticles in question (c") become incorporated in the next stage of 

 the growth of the crystal, or in some of the subsequent stages. 

 This Professor Dana has distinctly shown. What we have here 

 to observe is simply that, by reason of the reflex action above 

 noticed (p. 386), when new particles become united to c', c 1 , these 

 particles, c\ c\ will in fact exert a more energetic attraction late- 

 rally upon c"j c" ; and hence the union of c", c" with the crystal 

 may then be determined. If not, the next augmentation of the 

 attractive force attendant upon the union of the next set of mo- 

 lecules may determine this result. 



In order that a complete polyhedral crystal may be formed, it 

 is necessary that molecules on one or both sides of a plane in 



