LITHOPHYS.E. 417 



it often traverses the hollow spherulite and the gaping- or spaces are between 

 the layers, as in fig. 2. 



In lithophysse proper there are concentric shells of crystals and con- 

 centric spaces between them, as in fig. 3 of the same plate. The minute 

 feldspar prisms in the shell stand radially with respect to the center of the 

 lithophvsa?; the other minerals have no definite arrangement. Often the 

 tridymite occurs in minute pellets, dotting the shells. The concentric shells 

 correspond to the concentric bands of color and of varying composition 

 which characterize certain solid spherulites. They probably result from a 

 pulsation in the act of crystallization, such as has been observed in the 

 growth of crystals and spherulites of artificial salts when the latter grow 

 very rapidly. It is due to the lowering of the saturation of the surrounding 

 mother liquor, caused by the sudden liberation of heat in the act of crystalli- 

 zation, and to the rapid extraction of crystallizing molecules. The spasmodic 

 advance of the crystallization appears to have produced layers that were 

 more coherent than others, and the latter became the open spaces upon the 

 shrinkage of the partly crystallized magma before its final crystallization. 

 The position of the quartz, tridymite, and fayalite upon the surface of these 

 shells in some cases indicates that they formed after the spaces did. When 

 the groundmass is markedly banded the lithophysse are often modified, as 

 in fig. 4 of PI. LVII. Similar structures occur in hemispherical lithophysse, 

 as shown in figs. 5 and 6. The presence of these cross walls proves that 

 the shells are not the result of expanding gas bubbles, but must have been 

 formed by the contraction of the magma within the boundary of the litho- 

 physa at the time of its formation. In a very few cases there appear to be 

 evidences of a slight expansion, which is shown in the arching of the layers 

 of the groundmass over the lithophysa, as in fig. 7. But it is quite possible 

 that the curving of the layers may have antedated the formation of the 

 lithophysa, and have led to its formation. When we consider that the 

 known condensation of mass in the passage of anhydrous glasses of ortho- 

 clase and quartz into the crystal form is about 10 per cent in the first case 

 and 16 per cent in the second, and when we remember that hydrous glasses 

 possess still greater volume than the anhydrous ones, we are prepared to 

 understand the comparatively large cavities which often occur in these 

 forms of rapid crystallization. In the case of the dense spherulites, we 

 must assume that the condensation was more gradual, though rapid, and 

 MON xxxn, PT II 27 



