part of one side of the mountain, consisting of sloping beds of hard red sandstone and conglomerate, resting upon soft sandy layers, gave way. Thousands of tons of solid rock suddenly swept across the valley of Goldau, burying four villages, with about 500 of their inhabitants. In 1855 a Lass of debris, 3500 feet long, 1000 feet wide, and 600 feet high, slid into the valley of the Tiber, which, dammed back by the obstruction, overflowed the village of San Stefano to a depth of 50 feet, until drained off by a tunnel.
§3. Brooks and Rivers.
These will be considered under four aspects:-(1) their sources of supply, (2) their discharge, (3) their flow, and (4) their geological action. I. SOURCES OF SUPPLY.-Rivers are the natural drains of a land surface. They carry out to sea the surplus water after evaporation, and not water only, but a vast and almost incredible amount of material annually worn off the land. Their contents are derived partly from rain (including mist and dew) and melted snow, partly from springs. In a vast river system like that of the Mississippi, the area of drainage is so extensive as to embrace many different climates and varieties of rainfall, so that on the whole the amount of discharge, being in a great measure in- dependent of local variations in the weather, remains tolerably uniform. But in smaller rivers, such as those of Britain, whose basins lie in a region having the same general features of climate, the quantity of water is regulated by the local rainfall. A wet season swells the streams, a dry one diminishes them. Were rivers entirely dependent, however, upon direct supplies of rain, they would only flow in rainy seasons, and disappear in dry weather. This does not happen, because they derive a great deal of their water not directly from rain, but indirectly through the inter- mediate agency of springs. Hence they continue to flow even in very dry weather, because, though the superficial supplies have failed, the underground sources still continue available. In a long drought, however, the latter begin to fail, the surface springs ceasing first, and gradually drying up in their order of depth, until at last only deep-seated springs furnish a perhaps daily diminishing quantity of water. It is a matter of great economic as well as scientific interest to know how long any river would con- tinue to yield a certain amount of water during a prolonged drought. So far as we can tell, no rule could possibly be laid down for a generally applicable calculation, every area having its own peculiarities of underground drainage. Joseph Lucas gives some particulars which show what may happen in a chalk district. The river Wandle drains an area of 51 square miles of the chalk downs in the south- east of England. For eighteen months, from May 1858 to October 1859, as tested by gauging, there was very little absorption of rainfall over the drainage basin, and yet the minimum recorded flow of the Wandle was 10,000,000 gallons a-day, which, Mr Lucas says, represents not more than 4090 inch of rain absorbed on the 51 square miles of chalk. The rock is so saturated that it can continue to supply a large yield of water for eighteen months after it has ceased to receive supplies from the surface, or at least has received only very much diminished supplies.¹ Mr II. DISCHARGE.-As the natural drains of the land, rivers carry the surplus moisture out to sea. What proportion of the total rainfall is thus discharged by them is a question of great geological and industrial interest. From the very moment that water takes visible form as mist, cloud, dew, rain, snow, or hail, it is subject to evaporation. When it reaches the ground, or flows off into brooks, rivers, lakes, or the sea, it undergoes continual diminution from this cause. Hence in regions where rivers receive no tributaries, they grow smaller in volume as they move onward, till they some- times even disappear. Apart from temperature, the amount of evaporation is very largely regulated by the nature of the surface from which it takes place, one soil or rock differing from another, and all of them probably from a surface of water. There is no question in meteorology where full and detailed observations are more wanting than in the deter- mination of the relation of evaporation to rainfall and river discharge. During severe storms of rain, the water dis- charged over the land of course to a very large extent finds its way at once into brooks and rivers, where it causes floods, and whence it reaches the sea. Mr David Steven- son remarks that, acccording to different observations, the amount carried off in floods varies from 1 to 100 cubic feet per minute per acre. But though floods cannot be deemed exceptional phenomena, forming as they do a part of the regular system of water circulation over the land, they do not represent the ordinary proportions between rainfall and river discharge in such a climate as that of Britain, where the rainfall is not crowded into one season, but is spread more or less equally throughout the year. According to Beardmore's table,4 the Thames at Staines has a mean annual discharge of 32-40 cubic inches per minute per square mile, equal to a depth of 731 inches of rainfall run off, or less than a third of the total rainfall. The most carefully collected data at present available are probably those given by Humphreys and Abbot for the basin of the Mississippi and its tributaries as shown in the subjoined table :- Ohio River.... Missouri River... Upper Mississippi River.. Small tributaries. Arkansas and White River Red River..... Yazoo River... Ratio of Drainage to Rainfall.
0.15 0-20 0.90 St Francis River.. 0.90 Entire Mississippi, exclusive of Red River......... 0.25
0:24 0 15
0-24 0.90 Perhaps in Great Britain not more than a fourth part of the total moisture deposited on the land from the atmo- sphere is carried out to sea by streams. But this is a point on which, until far more facts have been gathered, no definite statement can be accepted as at all trustworthy. III. FLOW.-Rivers, in obedience to the law of gravita- tion, always move from a higher to a lower level. Where the channel of a river becomes vertical, or nearly so, a water-
1 Lucas, Horizontal Wells, London, 1874, pp. 40, 41. See also Braithwaite, "On the Rise and Fall of the Wandle," Minutes Proc. Inst. C.E., xx. 2 In the present state of our information it seems almost useless to state any of the results already obtained, so widely discrepant and irre- concilable are they. In some cases the evaporation is given as usually three times the rainfall! and that the evaporation always exceeded the rainfall was for many years the belief among the French hydraulic engineers. (See Annales des Ponts et Chaussées, 1850, p. 383.) Observations on a larger scale, and with greater precautions against the undue heating of the evaporator, have since shown, as might have been anticipated, that as a rule, save in exceptionally dry years, the eva- porations lower than the rainfall. As the average of ten years from 1860 to 1869 Mr Greaves found that at Lea Bridge the evaporation from a surface of water was 20 946, while the rainfall was 25-534 (Symons's British Rainfall for 1369, p. 162). But we need a vast accumulation of observations, taken in many different situations and exposures. in different rocks and soils, and at various heights above the sea. (For a notice of a method of trying the evaporation from soil, see British Rainfall, 1872, p. 206.) 3 Reclamation and Protection of Agricultural Land, Edin., 1871, p. 15. 4 Hydrology, p. 201. 5 In mountainous tracts having a large rainfall and a short descent to the sea, the proportion of water returned to the sea must be very much greater than this. Mr Bateman's observations for seven years in the Loch Katrine district gave a mean annual rainfall of 87 inches at the head of the lake, with an outflow equivalent to a depth of 81-70
inches of rain removed from the drainage basin of 71 square miles.