270 The American Geologist. April, i893 
glass were found in this interior, and hemispherical and concave 
fragments are not uncommon (Fig. 3,B). Very likeljMntact spheres 
were broken when the pellets were opened, for the material is ex- 
ceedingly brittle. 
The pellets give a peculiar and striking aspect to the rock, where 
the}^ occur. The question promptly suggests itself; how were they 
formed? Being found only in that part of the deposit, which is 
thoroughly ripple-bedded, it is near at hand to inquire, if wave 
motion may not have produced them. Examining a sand-beach 
•where the wind is throwing the waves against the shoi-e, it may be 
noticed that the diflferential motion back and forth in the water 
close to the bottom, often rolls and turns over sand gi'ains resting 
on the bottom. This rolling motion may be seen still better, if 
some lighter objects, such as small berries or seeds, or some soaked 
saw-dust, be thrown into the water. The wave current back and 
forth is quite noticeable half an inch above the bottom, but the 
water in immediate contact with the bottom is held bj' friction and. 
is but ver}' little affected by the motion. An object resting on 
the bottom will have its lower side held by this comparatively inert 
layer in the water as well as b}- the friction against the bottom, while 
the upper part will be exposed to the impinging force of the cur- 
rent just above. When the force of this current is strong enough, 
the object will turn over and roll. 
It seems that in water holding a great deal of sediment, an ob- 
ject thus set rolling, might gather up around itself minute part- 
icles adhering to it. Especially would this be the case, if the ob- 
ject has a rough surface. Such agglomerations would eventually 
become buried in the sediments, where they formed. 
The reader will remember that this volcanic dust is composed of 
thin flakes, which are the fragments of hollow glass spheres, 
formed b}' the expansion of gases in the magma thrown out by a 
volcanic eruption. The flakes (Fig. 3) are verj' irregular in shape. 
Some of the original small hollow glass globes are still entire and 
may be separated from the rest of the dust by throwing it into 
water. The broken flakes will then sink, but the little entire 
spheres will float. When the dust fell a number of the entire 
hollow spheres no doubt floated on the water for some time, as 
they will do 3'et. The angular character of some of the glassy- 
fragments must have caused them to become entangled with such 
floating spheres and thus to sink them. Most such clusters would 
