A lake with an outlet has its level determined by the height of the outlet. Great Salt Lake, having no outlet, has its level determined by the relation of evaporation to inflow. On one hand the drainage of a great basin pours into it a continuous though variable tribute; on the other, there is a continuous absorption of its water by the atmosphere above it. The inflow is greatest in the spring time, while the snows are melting in the mountains, and least in the autumn after the melting has ceased, but before the cooling of the air has greatly checked evaporation on the uplands. The lake evaporation is greatest in summer, while the air is warm, and least in winter. Through the winter and spring the inflow exceeds the evaporation, and the lake rises. In the latter part of the summer and in autumn the loss is greater than the gain, and the lake falls. The maximum occurs in June or July, and the minimum probably in November. The difference between the two, or the height of the annual tide, is about 20 inches. But it rarely happens that the annual evaporation is precisely equal to the annual inflow, and each year the lake gains or loses an amount which depends upon the climate of the year. If the air which crosses the drainage basin of the lake in any year is unusually moist, there is a twofold tendency to raise the mean level. On one hand there is a greater precipitation, whereby the inflow is increased, and on the other hand there is a less evaporation. So, too, if the air is unusually dry, the inflow is correspondingly small, the loss by evaporation is correspondingly great, and the contents of the lake diminish. This annual gain or loss is an expression, and a very delicate expression, of the mean annual humidity of a large district of country, and as such is more trustworthy than any result which might be derived from local observations with psychrometer and rain gauge. A succession of relatively dry years causes a progressive fall of the lake, and a succession of moist years a progressive rise. As the water falls it retires from its shore, and the slopes being exceedingly gentle the area of the lake is rapidly contracted. The surface for evaporation diminishes and its ratio to the inflow becomes less. As the water rises the surface of the lake rapidly increases, and the ratio of evaporation to inflow becomes greater. In this way a limit is set to the oscillation of the lake as dependent on the ordinary fluctuations of climate, and the cumulation of results is prevented. Whenever the variation of the water level from its mean position becomes great, the resistance to its further advance in that direction becomes proportionally great. For the convenience of a name, I shall speak of this oscillation of the lake as the _limited oscillation_. It depends on an oscillation of climate which is universally experienced, but which has not been found to exhibit either periodicity, or synchrony over large areas, or other features of regularity.