knives, etc. is desired, yet in mosaics and work enriched by a display of iridescent tints the nacre of abalone shells stands preeminent. Inlaid work is so universally used that an enumeration of articles ornamented in this way is unnecessary, but mention may be made of one use of these shells in lacquer that to an American or European may seem unique; its use in a "pillow end." When we think of a pillow we imagine a billowy roll all done up in white, but, a Japanese or Korean has a very different idea. In the Korean collection in the U. S. National Museum are some small pillows and the following description is given of the ends of two of them:8 "Pillow end (Be-ga-mo). Circular piece of wood, lacquered, incrusted with Haliotis shell. Figures represent a tiger under a pine tree; along the border is a band of arabesque." "Pillow end (Ja-ga-be-ga-mo). Disk of wood fastened in the end of a cylindrical pillow case, in black lacquer with Haliotis shell. Subject, the great dragon rising from the sea into the sky in the spring season." In describing these pillow ends Mr. Walter Hough says: "The Korean pillow is a cylindrical case stuffed with hair or rice straw. It has ornamented ends. The first one mentioned is 8 inches in diameter, but is not part of a regular pillow,' being used as a 'arm-rest.' The second one is 8 inches in diameter."

As a medium for trade among the Aborigines of North America, abalones have been highly esteemed both for their beauty and importance when used as shell money. The shells in the latter case being cut "into oblong strips from one to two inches in length, according to the curvature of the shell, and about as third as broad as long." These were strung on a string and were used both as money and ornaments. Dr. Robert E. C. Stearns, Adjunct Curator of Molluscs in the National Museum, has written a comprehensive monograph upon the use of shells by the Indians, entitled "Ethno-Conchology, a study of Primitive Money," and in it is figured money made from abalones, which the Indians termed "Uhllo." In the recent excavations at the old historic town of *Report of the U. S. National Museum, 1891, page 465.

Pachacamac, near Lima, Peru, squares of mother-of-pearl were found in the graves of the Incas. These squares are only half the length of those figured in Dr. Stearns' paper. The pieces look like the nacre of abalones and each square has two holes drilled in it. As the graves, or burial place of Pachacamac is supposed to be over four hundred years old, these shell pieces are very interesting, revealing also the fact that the Incas considered shell ornaments valuable enough to be buried with their bodies. As these strips of solid silver, done up in a loosely woven cloth, were found in a mummy's hand, the pieces of shell were evidently not used as money, the silver having been cut for that purpose.

Dr. Stearns instances the purchasing power of an abalone from the fact that in New Mexico a horse had been traded for a shell. I was relating this incident to a friend who had spent some years with the Pueblos in New Mexico, and my friend said that that was not surprising, as, when she first went to New Mexico, some years ago, her brother bought her a good Mexican horse for $6.00, and the Indians were always as glad to receive attractive shells as money. This would not be a very extravagant price for an Indian to pay for a fine Haliotis, as a shell dealer once listed to me H. fulgens as high as $10.00. Whether any conchologist paid such price is unknown to me, but, a red abolone, when decorticated, has sold in Los Angeles for $5.00, but it was a large specimen and beautifully polished.10 Like other commodities abalone shells are variable in price according to the demand, as well as quality.

"In the private collection of C. F. Lummus, Los Angeles, Cal.

10It is related that as high a price as $25.00 has been asked for an abalone having a peculiar muscular impression outlined in the interior of the shell.

THE DURATION OF NIAGARA FALLS.

BY DR. J. W. SPENCER.1

For the past century Niagara Falls has been considered a time measurer, but its greatest interest has risen since the growth of our knowledge of the Ice Age on acconnt of the expectation that in some way it can be made to tell something of the date of that period and indirectly of the advent of man, or his restrictions on account of the glacial conditions. The paper of which this is an abstract was primarily a physical study, setting forth the changing episodes in the history of the falls, and computing the age of the river, but leaving to others the application of the results in the question of early man.

The method of determining the age of the falls is the application of the mechanics of the river to the various conditions during the changing episodes of its history, in a large measure discovered by the author during the last fifteen years. The investigation differs from those of other writers who have simply divided the length of the chasm, excavated by the retreating falls, by the imagined or measured rate of the recession of the cataract. At a glance, even the most superficial reader can understand that if the height of the cataract be first reduced to one-half, and then again doubled, or if the volume of the river be reduced to one-fourth, such variations are bound to produce as great changes in the rate of recession as are indicated by the mechanical laws; and that if the conditions have not always remained constant, then the present rate of retreat has not always obtained-sometimes slower and sometimes faster. It is this question that the paper considers for the first time. In the much written, but, until recently almost unknown, history of Niagara River, we find that an approximately correct estimate of the age of the falls was made half a century ago by Lyell, upon a conjecture of the rate of re

1Abstract of a paper read before the Am. As. Ad. Science at Brooklyn, August, 1894.

cession now known to be wholly erroneous. Again, within the last eight years, there have been several writers who have been using corrected cöefficients of retreat, still their results are more inaccurate than the guesses, as to the age of the falls, made a hundred years ago, yet they may be said to have approximated the truth within their observations, but the observations have become enlarged.

A hundred years ago, Andrew Ellicott estimated the age of the falls at 55,000 years. Forty years later, Bakewell made the falls about 12,000 years old. Over fifty years ago, Lyell conjectured the age at 35,000 years, and this estimate was commonly accepted until about a decade ago. The foundations for the measurements of the retreat of the cataract were laid by Professor James Hall, when he made the first preserved instrumental survey of the cataract in 1842. Since then, measurements have been repeated in 1875 by the Lake Survey; in 1886 by Professor W. S. Woodward, and in 1890 by Mr. Aug. S. Kibbe. From these surveys the mean rate of modern recession of the falls is found much more rapid than was formerly supposed, as it amounts to 4.175 feet a year, and if the history of the falls had been uniform, then the age would have been only 9,000 years-not so different from the guess of half a dozen years ago, which took the maximum medial retreat of the cataract, and made the age only 7,000 years. Had the gentlemen taken the mean rate as then known, which the scientific methods dictated and since supported by the action of the river, they should have made the age of the falls 11,000 years, near which estimate some did. This point is noticed on account of many secondary writers finding the number 7,000 years as agreeable to their theories.

Owing to some structural variations, I have taken 3.75 feet a year as the mean rate to be adopted for the retreat of the falls mechanically applied to the different conditions of the river. These have been occasioned by the changing heights of the falls and the volume of the water. With regard to the latter point, it has been found that for three-fourths of the duration of the river, the drainage of Lake Huron and the upper lakes was by way of the Ottawa River, and not by way

3 11

of Lake Erie and the Niagara. Under these conditions only fr of the present discharge of the Niagara River cascaded over the falls. The episodes of the river are as follows: First episode water descending 200 feet, volume of the present (when the falls was of about the magnitude of the present American cataract), chasm excavated (as shown by the position of terraces) 11,000 feet; time required, 17,200 years. Second episode: descent of the river in a series of three cascades aggregating 420 feet at first with only the Erie drainage (during the recession of 3,000 feet) and afterwards the present volume of water (when the recession amounted to 7,000 feet) duration 10,000 years. Third episode: river descending 420 feet in one cascade with the present volume; time required, for the recession of 4,000 feet, only 800 years. Fourth episode was somewhat complicated, with the water mostly descending 320 feet, and during this condition the falls have receded 11,500 feet, and required a period of 3,000 years. Thus the age of the falls has been computed at 31,000 years. But at the beginning, the river flowed from lake to lake without a falls, and this time has been taken as 1,000 years; accordingly, the age of the river is computed at 32,000 years. The record of the changing levels may be seen in the deserted beaches now high above the lakes which have already been described in scientific journals. The investigations doubtless contain some errors which may be corrected in the future, but in the history of the lakes the present computations are very strongly confirmed by much cumulative evidence so that the present results appear to be approximately correct. It is further estimated that with the earth movements continuing as at present, the end of the falls will be effected by the change of the drainage from the Niagara River to the Mississippi, by way of Chicago, owing to the rise of the eastern rim of the Erie basin above the barrier now separating the lake waters from the Mexican drainage. With the present rate of elevation continuing, the future life of the river ought to be 5,000 or 6,000 years.

In regard to the relation of Niagara River to the Ice Age, I estimate that the lake epoch commenced from 48,000 to