hills in the Maumee valley. Finally the blockade was raised in the St. Lawrence valley, the outlet of Ontario was shifted from Rome to the Thousand Islands, and its water level was drawn down five hundred feet. During the Rome epoch of its history Ontario's area was 60 per cent. greater than now; it began the Thousand-Island epoch with an area 30 per cent. smaller than

now.

While yet the glacier was present and the navy of Ontario was a fleet of icebergs, the depressed land at the north had begun to rise again. When the glacier was quite gone the reflux was rapid, the land soon reached a more stable position, and the lakes acquired their present dimensions. Had the oscillation received no check, our hydrography and avenues of commerce might have been very different; a further tilting of the land to the extent of three inches in each mile would send a great river from Chicago to the Mississippi, reverse the current in the Detroit, stop Niagara Falls, and rob the upper St. Lawrence of seven-eighths of its water.

Has the oscillation ceased? Is Niagara destined to run dry? These are questions hard to answer for the remote future into which science fain would peer, but less difficult as concerns those few generations of posterity to which our ambitions and sympathies extend. It is one of the inductions of geology that absolute stability is a myth, and all parts of the earth's crust continually undergo changes of level. There is no reason to believe that the lake district is an exception to this law, but whatever movements may be there in progress are so slow that they have not been detected, and their tendency is unknown. In our use of the lake harbors we have observed no changes requiring earth movements for their explanation, and this negative testimony, so far as it goes, shows present stability. That which the waves have done to the present coasts, in the cutting back of cliffs and the building of spits, is a work of many centuries, during which the water level must have remained nearly constant; and the practical stability thus shown for the immediate past is a guarantee for the immediate future.

There is no question that changes of other kinds are in progress Storm waves and storm currents are eating away the coasts

and spreading the fine débris over the lake bottoms, where it mingles with the muddy tribute brought by flooded rivers; the St. Clair River is feebly scouring its channel and building its delta; the Falls of Niagara are gnawing back toward Lake Erie at the rate of four or five hundred feet in a century; and with infinite slowness the Ste. Marie, Detroit, and St. Lawrence Rivers are deepening their rocky beds. In a future geologic age all the lakes that survive the erosion of outlets will have succumbed to filling by alluvial mud, and the reign of running water will once more be established.

But the lake basins are so capacious that we become aware of their slow filling only by observing the discoloration of the water in times of freshet and of storm. The scour of the St. Clair can do no more than reduce the level of Lake Huron to that of Lake St. Clair, a difference of two or three feet; and as the reduction proceeds, its rate, now exceedingly slow, must continually diminish. If the recession of Niagara Falls were to continue at its present rate, Lake Erie would be tapped in two hundred centuries; but the rate is determined by the geologic structure, and that structure changes between Goat Island and Buffalo in such a way as to retard the work of erosion.

All these processes are too slow to affect our hopes or fears concerning the immediate future, and for our posterity in the year 20,000 we have no solicitude. The men who shall watch the draining of Lake Erie-or who, perchance, shall find it worth their while to prevent it-will as far surpass us in powers and resources as we surpass the men who watched the lake's creation. For all practical purposes our inland seas are perma nent and their basins stable. The only modifications that affect our economy are those wrought by the waves upon their coasts.

Nevertheless, their stability is sometimes called in question. Their levels are not absolutely constant, but oscillate under various influences about a mean position, and when they are unusually low the "oldest inhabitant" is interviewed, and is reported to declare that the like was never seen before. Then some theory of permanent change is promulgated and the sensa tion has its day. While yet the newspaper discussion of the recent lowering of water levels is fresh in memory it will not be

amiss to recite briefly the conditions on which such changes depend.

About each of the lakes is a district of land draining toward it. A portion of the rain and snow falling upon this land is returned to the atmosphere by evaporation from the soil, and a larger portion is returned by evaporation from the surfaces of plants. The remainder flows to the lake and tends to raise its level. Its level is also raised by the rain and snow falling upon it. On the other hand, its level is lowered by evaporation, and is lowered by the discharge through the outflowing river. In the long run the supply from inflow and rain is balanced by the loss through evaporation and outflow, and so in a general way the lake altitude is constant; but in detail it is inconstant, oscillating about its average position.

The additions to the lake by rain are not uniform through the year, but are usually greater in summer. The additions from tributary streams are still less uniform, being smallest in winter, when precipitation takes the form of snow, and largest in spring, while the snows are melting. The loss from evaporation is likewise unequal, varying with the temperatures of air and water, with the dryness of the air, and with the velocity of the wind, and being usually greatest in summer and autumn. Thus supply and loss are not balanced in detail; at some seasons there is a net gain and the lake surface rises, at others there is a net loss and it falls, and the rise and fall together constitute an annual oscillation.

A second difference depends on the variation of weather from year to year. In some years more rain and snow fall, in others less, and there is a similar fluctuation in the atmospheric conditions affecting evaporation. When the rainfall is greater than usual or the evaporation less, the lake rises; when the rainfall is small or the evaporation great, the lake falls. A succession of wet years produces exceptionally high water, a succession of dry years extremely low water. But there is a limit to such cumulative effects, for when the lake is high its outflow is more rapid than when it is low, and an automatic check is thus furnished.

Thirty years ago Colonel Charles Whittlesey compiled all available data regarding the fluctuations of the lakes, and was

able to publish an account of the more important changes of the lower lakes between the years 1838 and 1857, together with a few data concerning exceptional phenomena in earlier years. In 1859 the United States engineers began systematic gauge-readings, and their work is still continued. The following table is based on their records, and shows the ordinary range of fluctuations. Michigan and Huron are here treated as one lake because their waters communicate freely through a strait.

Michigan-Huron.. 1860-87 July or Aug. Jan. or Feb.

April.

1.3

3.0

The highest water known occurred in 1838, when MichiganHuron rose 26 inches above ordinary high stage, and Erie and Ontario 18 inches. The lowest water known was in 1819, when Erie fell about 3 feet below its usual plane.

The present low water is the sequel of last summer's drouth. The Signal-Service records indicated that the lake region received in the year 1887 only about 26 inches of rainfall instead of its usual quota of 33 inches. If the evaporation and the discharge remained constant, the lakes should fall 7 inches by reason of the defect of aqueous precipitation on their surfaces, and about as much more by reason of the defect of inflow; but, taking the average for all the lakes, the actual fall from the low stage of 1887 to the low stage of 1888 has been only 7 inches. The variation of rainfall was, therefore, great enough to account for the variation of lake surface. That it was more than sufficient is probably explained by the coolness of the autumn, which tended to diminish evaporation. In Superior the low water of last February reached 5 inches below the level of the average low stage, a depression exceeded but twice in seventeen years.

In 1879 and 1880 the water was 3 inches lower. In Ontario, the lake most affected, the low water of 1888 was 6 inches below the average, but this record has been exceeded six times in the last twenty-eight years. 1868 marks 10 inches below, 1872 16 inches, 1873 14 inches, 1875 (February) 13 inches, 1875 (December) 7 inches, and 1881 11 inches. In Michigan-Huron the recent low water was but 2 inches below the average, and in Erie but one inch. If our inland commerce has need to be assured of the continued fidelity of its "unsubsidized ally," it can find comfort in the contemplation of these figures.

The oscillations described affect an entire lake uniformly. There are others that affect its parts differently, the water rising in one place while falling in another. The most powerful cause of such displacement of level is the wind, which, driving the surface water before it, heaps it up against the lee shore; and the greatest effects are seen in Erie, whose shallowness interferes with the adjustment of levels by means of a return current beneath. A gale blowing in the direction of the lake's length has been known to raise the level seven or eight feet at one end and depress it an equal amount at the other.

Oscillations of a second kind are caused by inequalities and variations of atmospheric pressure. When the air presses unequally on different parts of a lake an equilibrium is reached by a depression of the water surface under the heavier column and its elevation under the lighter. If the air pressures are rapidly shifted, as in the case of thunder-storms and tornadoes, rhythmic undulations are produced analogous to those from the dropping of a pebble in still water, and traveling like them to remote shores. The rhythmic period is usually measured in minutes and the height of the undulation in inches, but waves of this class sometimes equal the largest generated by wind. The passage over Lake Michigan of a broad wave of barometric change sets the water to swaying from side to side as we sometimes see it in a hand basin; but the greater body has a longer period, advancing and receding only eleven times in twenty-four hours.

Third and last are the tides, which ebb and flow in lunar and solar cycles as regularly here as on the ocean, but are unheeded