layers from above downward, for it is evident that as the direction of the wind changed, the ripples were also turned. The deciphering of this record must be made backward. The bottom layers were deposited first, and the excavation must begin on top. Otherwise the record is 
Fig. 5.—Ripple Marks in the Volcanic Dust. Reduced to 1/4 diameter. as perfect as if it had been taken down by an instrument when the shower occurred. It may be only local in its significance, for it shows the direction of the wind at this particular place alone. The wind may have been somewhat deflected from the general direction by local topographic peculiarities, though these appear to have been of small importance. In any case, the old legend is quite interesting to read, being, I believe, the only geological record ever found of the passing of a cyclone over the United States.
In the lowermost foot of the deposit no ripple marks can be seen. But there appear some marks of sedges and other vegetation, and these are inclined to the west, as if the plants had been bent by an east wind. Just above the height to which the imprints of the vegetation extend, ripple marks begin to appear, running on a northeast-southwest course. They were made by a southeast wind, for their northwest slopes are the steeper. A little above this height some peculiar small elevations appear on one of the bedding planes, and slightly raised ridges run for a short distance to the northeast from
Fig. 6.—Peculiar Elevations caused by a Current from the Southwest to the Northeast. Reduced to 1/2 diameter.
each elevation, vanishing in the same direction (Fig. 6). A southwesterly current was unmistakably obstructed by the little elevations, and left the small trails of dust in their lee. Six inches higher up the wind comes more from the south, and for the next foot the ripples continue to gradually turn still more in the same direction so as to at last record a due south wind. At this point it suddenly changed and
