196 SIR DAVID BREWSTER ON THE STRUCTURE AND 



another, and a long groove is excavated, the bottom of which is formed of spherical 

 cavities of different depths and breadths, depending on the distances of the centres 

 of decomposition, and the rapidity with which the decompositions have taken place. 

 The form of the grooves is thus very irregular, being wide at one place and naiTow 

 at another, and occasionally not rectilineal. When the centres of decomposition 

 are very near each other and equidistant, the grooves are very narrow and 

 shallow, and form straight lines of uniform breadth. 



These grooves or lines of cavities lie in different directions, frequently crossing 

 one another at various angles. In some specimens they are perfectly parallel, and 

 are often not more than the 500th or 1000th of an inch in breadth, having the 

 appearance of lines of uniform thickness. 



When the cavities are of considerable size, and nearly or perfectly spherical, 

 they sometimes present remarkable phenomena. Round one or more points of 

 the spherical films or cavities, a fresh decomposition has taken place, and formed 

 smaller spherical cavities, sometimes almost touching one another, as in Plate X. 

 fig. 4, and Plate XI. fig. 6 ; and occasionally so numerous, that I have found from 

 fifteen to twenty formed upon the same cavity. In some cases the decomposition 

 has produced shallow cavities, which, from interfering with one another, have 

 become polygonal. In some specimens this secondary decomposition has taken 

 place at such a great number of points, that the infinitely small cavities which it 

 has formed appear like specks of black powder, making the cavity more or less 

 opaque, according to the distance of the specks, or the number of films of which 

 the cavity is composed. 



When the cavities do not interfere with one another, they are occasionally 

 ellipsoidal, and egg-shaped, though formed from one centre of decomposition. 



Specimens of these structures are given in Plate X. fig. 3, and Plate XL figs. 

 7 and 12, seen by ordinary transmitted light. In Plate X. fig. 5, are shown 

 circular cavities, and cavities in lines of uniform and variable breadth. The com- 

 pound film consists of many elementary films, but only three or four are shown 

 in the figures, as they alone are visible in the specimen. If we remove with a 

 lancet those films in succession, we shall find that each film contains all the cavities 

 existing in the compound one, — the spherical or polygonal cavity, whether deep 

 or shallow, being joined to, or forming part of the film. When examined on 

 one side all the cavities are convex, as shown in Plate XI. fig. II; and when 

 examined on the other they are concave, as shown in Plate XI. fig. 13. 



In comparing this structure of the films with the process of decomposition 

 shown in Plate X. fig. 1, it is not easy to understand how the concave or convex 

 parts of the film are united to and form part of the general film. The process of 

 decomposition having begun at one point, where the cohesion of the particles is the 

 feeblest, it advances, forming round that point a number of spherical films, before 

 the decomposition has commenced on the surrounding glass. When the surface 



