84 F. B. TAYLOR—ORIGIN OF GORGE OF WHIRLPOOL RAPIDS. 
of the present American fall, but the conditions surrounding this cataract 
are so anomalous, permitting it to remain almost entirely stationary, that 
although it is apparently the best we have, it seems to be almost useless 
as a criterion.* 
SUMMARY 
For as much of the Niagara gorge as is brought under discussion in 
this paper, the correlations between the gorge and the lakes seem clear 
and complete, point for point and episode for episode. They may be 
summarized briefly as follows : 
1. The episode of the Middle Great gorge, extending from Wintergreen 
flat up to the upper side of the Eddy basin, is the correlative of the latter 
part of the episode of lake Algonquin—that part which comes after 
the closing of its Trent Valley outlet, supposing the lake to have drained 
for a time in that direction. 
2. The place of the sudden contraction of the gorge at the upper side 
of the Eddy basin is the correlative of the breaking of the ice-dam in the 
Ottawa valley and the opening of northeastward drainage—the end of 
lake Algonquin. 
3. The episode of the gorge of the Whirlpool rapids is the correlative 
of the episode of Nipissing Great lake with the Nipissing-Mattawa river 
at its outlet—episode of the Champlain marine submergence. 
4. The place of expansion at the lower end of the Upper Great gorge 
above the railroad bridges is the correlative of the change of outlet of 
the upper lakes from North Bay to Port Huron—the beginning of the 
Champlain uplift. 
5. The episode of the Upper Great gorge is the correlative of the 
modern or post-Champlain episode of the upper Great lakes. 
*So long as the American fall was taken as a unit of measure, the writer’s estimate of the dura- 
tion of postglacial time was expressed in large figures with a very wide range between extremes. 
No limit short of 100,000 years or more was recognized as demonstrable, but an estimate of 60,000 
to 70,000 years was -considered most satisfactory. New evidence haying some bearing on this 
problem was found in 1896 and 1897 in the drowned lower courses of the tributaries of the Saint 
Clair and Detroit rivers and lake Saint Clair. (F. B. Taylor: ‘Some features of the recent geol- 
ogy around Detroit,’ (Abstract) Proc. Am. Assoc. Ady. Sci., 1897.) The beds of all the tributaries 
have evidently been deepened ata time when the main river beds were abandoned, the upper lakes 
discharging at that time over Nipissing pass, and then have been drowned or flooded afterwards 
when the discharge of the upper lakes returned to this course. The drowned stream beds indicate 
a relatively long duration of timeat lower base level, but they do not appear to support so large 
figures as those mentioned aboye. Revising our conclusions in the light of recent advances, it 
may be said, tentatively, that 50,000 years may be regarded as an approximate extreme limit for 
the making of the whole gorge of Niagara, but that it may have been as short as the estimates of 
Lyell (35,000 years), or Spencer (32,000 years). It ought to be distinctly recognized, however, that 
many of the elements of the Niagara time problem are, in the very nature of things, so uncertain 
in their values that no time estimate pretending to accuracy within narrow limits can be trust- 
worthy. This is the more unfortunate because the Niagara gorge is by far the best single datum 
for estimating the duration of postglacial time that has yet been discovered, 
