456 
IDDINGS. 
are locally abundant in very small spaces within the body 
of the rock, and not along a contact face of it. The crys¬ 
tallization of the tourmaline at least involves the presence 
of a small amount of boron and fluorine within the magma 
before its final solidification; but they were probably present 
in extremely small amounts and only locally. 
While the occurrence of the tourmaline like that of the 
fayalite may be referred to the category of “ fumarole action,” 
still this is only correct when the term is so defined as to 
include any mineralizing influence which heated vapors may 
have upon crystallization. It would thus include their 
primary action within fused magmas, as well as their sec¬ 
ondary action on solidified ones. The effect of heated vapors 
which permeate the rocks in many places in the Yellowstone 
National Park is distinctly a destructive or metamorphosing 
process; and all such fumarole action is plainly secondary 
in the sense that it affects changes in the crystalline character 
of rocks already solidified. It would seem advisable, there¬ 
fore, in order to avoid confusion, to use some other term for 
the primary mineralizing influence of absorbed vapors upon 
the crystallization of molten magmas. 
The second kind of spherulites occurring in this rock, 
which were described in the paper on Obsidian Cliff under 
the head of porous spherulites (p. 278), are distinguished by 
being composed of more or less distinct rays of feldspar, 
which are generally branched, and a cementing material of 
tridymite, with numerous hollow spaces or gas cavities. The 
branching feldspars may also lie in an isotropic, base, which 
appears to be glass. The arborescent feldspar growth may 
form a complete sphere or only a plume-shaped growth, or 
it may even resemble the stem and branches of a shrub. 
The optical investigation of these feldspar rays shows them 
in some cases to consist of many small stout prisms of feld¬ 
spar grown together with their longer axes parallel and 
forming long crooked rods in the direction of this axis. 
In thin section these rods or rays are partly positive and 
partly negative-—that is, the axis of elasticity, which is 
