But, to prepare the way for a more definite settlement of the question, the New York Geological Survey, under Professor Hall's direction, had a careful trigonometric survey of the Horseshoe Fall made in 1842, erecting monuments at the points at which their angles were taken, so that, after a sufficient lapse of time, the actual rate of recession could be more accurately determined. In 1886 Mr. Woodward, of the United States Geological Survey, made a new survey, and found that the actual amount of recession in the center of the Horseshoe Fall had proceeded at an average rate of about five feet per annum. The subject was thoroughly discussed by Drs. Pohlman and Gilbert, at the Buffalo meeting of the American Association in 1886, when it was proved, to the satisfaction of every one, that, if the supply of water had been constant throughout its history, the whole work of eroding the gorge from Lewiston to the Falls would have been accomplished, at the present rate of recession, in about seven thousand years. But the question was immediately raised, Has the supply of water in Niagara River been constant? It was my privilege, in the autumn of 1892 (see Bulletin of the Geological Society of America, vol. iv, pp. 421-427), to bring forth the first positive evidence that the water pouring over Niagara had for a time been diverted, having been turned through Lake Nipissing down the valley of the Mattawa into the Ottawa River, following nearly the line of Champlain's old trail and of the present Canadian Pacific Railroad. The correctness of this inference has been abundantly confirmed by subsequent investigations of Mr. F. B. Taylor and Dr. Robert Bell.[A] The occasion of this diversion of the drainage of the Great Lakes from the Niagara through the Ottawa Valley was the well-known northerly subsidence of the land in Canada at the close of the Glacial period. When the ice melted off from the lower part of the Ottawa Valley the land stood five hundred feet lower than it does now, but the extent of this subsidence diminished both to the south and the west, making it difficult to estimate just how great it was at the Nipissing outlet. A subsidence of one hundred feet at that point, however, would now divert the waters into the Ottawa River. That it actually was so diverted is shown both by converging high-level shore lines at the head of the Mattawa Valley and by the immense delta deposits at its junction with the Ottawa, to which attention was first called in my paper referred to above. [Footnote A: See article by Mr. Taylor on The Scoured Bowlders of the Mattawa Valley, in the American Journal of Science, March, 1897, pp. 208-218.]