That the dust was produced in this way is quite evident from other circumstances. If a handful from the dust of this place be thrown into water and gently stirred, it nearly all will settle after a while. But some rather large particles remain floating on the surface. If these are removed and examined under the microscope, they are seen to be hollow spheres (Fig. 2, _b_). These are some of the original bubbles that never burst, either because they contained too little steam or else because the steam was cooled before it had time to break the walls open. It is evident that not every droplet of the molten magma would form a single sphere, but that many also would swell up into a compound frothlike mass of pumice. A few such pieces may sometimes be observed in the deposit at this place, and that many more were made and broken is evident from the great number of glass fragments which have riblike edges on their flat sides (Fig. 3, _a_). The nature of the force which caused the eruption may thus be understood from the study of one little grain of the dust, but much more extended observations are needed in order to make out the place where the great convulsion took place. It will, perhaps, never be known what particular volcanic vent was the source of this ash. Different deposits may have come from different places. But it seems possible that it all came from the same eruption. There can be no doubt that the volcanic disturbances occurred to the west of the Great Plains. No recent extinct volcanoes are found in any other direction. This conclusion is corroborated by the fact that the dust is finer in eastern localities and coarser nearer the Rocky Mountains. In a bed near Golden, in Colorado, seventy-three per cent, by weight, of the dust consists of particles measuring from one fourth to one thirty-second of a millimetre, while some from Orleans, in Nebraska, contains seventy-four per cent of particles measuring from one sixteenth to one sixty-fourth of a millimetre in diameter. Still finer material comes from the bluffs of the Missouri River near Omaha. Evidently the coarser particles would settle first, and if the dust is finer toward the east, it must be because the wind which brought it blew from the west. Most likely the eruption occurred somewhere in Colorado or in New Mexico.