seasons at which each half of its tributaries is flooded; the one in January, the other, which is the highest and the most dreaded by planters, in the month of June.*

The great affluents of the Amazon, north of the equator, are flooded when the sun has passed that line and is advancing towards the tropic of Cancer, and the southern affluents during his passage towards the tropic of Capricorn.† Thus an equilibrium is established, and while the main stream has its own inundation during the spring and early summer, when it overspreads the country, sometimes for a hundred miles, at all times it retains its majesty, and rolls towards the Atlantic a vast body of water miles in width.

But high as the floods of these great rivers rise, their inundations can never be relatively as great as those of smaller streams, over the whole of whose basin some great storm of rain may burst at once. Elisée Reclus gives from Marchegay's "Annales des Ponts et Chaussées" an account of an extraordinary inundation of the Ardeche, in which that little river, not greater than a second rate tributary of the Tiber, rose in 1837 at the bridge of Gournier to the height of twenty metres and two-fifths of a metre, or about seventy English feet. On the other hand, the highest recorded flood of the Garonne,‡ rose only to thirteen metres or forty-two feet and a half, and the greatest height ever attained by the Seine was eighteen metres or thirty-two and a half English feet.

The basin of the Tiber, whose superficies exceeds that of the Thames by two thousand square miles, is large in proportion to the length of the main stream, and owing to the compactness of its form, which approaches in shape to a semicircle, of which the centre is the confluence of the Nera, no one point is so far removed from another as to render it extremely improbable that a great fall of rain should occasionally extend over the whole of its surface, and swell simultaneously its tributary streams.

* Elisée Reclus.

+ Idem.

Raulin Geographie Girondine, quoted by Elisée Reclus.

2. RAINFALL.

The rainfall within the basin of a river, and its distribution among the months of the year, is the second condition on which the occurrence of floods depends. How much depends upon the distribution of the rain appears from the case of Australia, where frequent droughts, during which the sheep perish by thousands, alternate with floods which are equally destructive to the crops and the cattle. Yet the annual amount of rain, both for Melbourne and Sidney, if uniformly spread over the year, would clothe the fields in perpetual verdure, and maintain a full and perennial flow in the rivers.*

The average rainfall for Rome is a little above thirty-one inches, the extremes being forty-three and nineteen inches, neglecting decimals; and nearly the whole of this quantity falls during the nine months of the year when its effect is the greatest. The average for Florence is given at forty-two inches. If this be correct, the difference must be owing to the greater proximity of Florence to the Apennines. If fortytwo inches be the average at the foot of the Apennines, and the extremes bear the same relation to the average as at Rome, we shall have about sixty inches, as the quantity which in exceptional years may fall over the whole region through which the tributaries of the Tiber flow.

From the source of the Tiber on the north to that of the Salto on the south, the distance, as a bird would fly, as I have said, is one hundred and forty miles. For this distance all the rain and melted snow from the western slopes of the Apennines must fall within the basin of the Tiber. If to this we add the contributions of the western tributaries, we shall have an explanation of the inundations of the river, as far as they depend upon the depth of rain which falls.

But how much of this rain finds its way into the river? This depends upon the permeability or absorbent qualities of

* At Melbourne there fell in the year 1870 sixty inches of rain, and in the three months of January, February and March, 1871, sixteen inches. But this must be above the average, which I have not yet had an opportunity of ascertaining.

the soil, and upon the number of tributaries which intersect the basin of the river.

3. PERMEABILITY OF THE SOIL.

Of all soils the most permeable is pure silicious sand. It not only furnishes no water to the streams which traverse it, but abstracts a portion of their volume, and if the sandy district be extensive, ends by swallowing them up entirely. In this way rivers of considerable size are lost in the sandy deserts of Africa and Central Asia. Next in order of permeability comes the chalk, which is characterised, as everyone may have observed, by extreme dryness. No springs originate in strata of chalk, and no rivulets flow through chalky valleys. All the rain which falls upon the surface percolates through the soil, until it is stopped by some impervious stratum, when it either forms reservoirs in the earth or issues as a spring at a lower level. The same description will apply to the four inferior members of the Oolitic series, which are nearly as absorbent as the chalk. Vallès observes that in a district of two thousand square kilometres of such a formation, there was no trace of streams, all the rain which fell disappearing in the earth.

In a letter from the engineer Belgrand to the Geological Society of France, it is stated that on the fifteenth and sixteenth of October the enormous quantity of 1655 metres, or about six and a half inches of rain fell in the upper part of the basins both of the Loire and the Seine, rather more in the basin of the Seine than that of the Loire. Yet the Loire, which, in the upper part of its course, flows through the impermeable granite, was swollen to a great height, and caused losses which, according to the Moniteur of third June, 1847, amounted to forty millions of francs. The rise of the Seine, on the other hand, was so trifling that it attracted no notice.

The Medway affords another illustration of the connexion between the floods of a river and the nature of the soil through which it flows. The Medway is a river of a short course, but so sluggish that its floods take thirty hours to

descend the stream to Yalding, which is distant not more than the same number of miles from its source. Its tributaries all flow through the Wealden clay and Hastings sand, both impermeable strata. As soon as the clay is saturated in autumn, the rain runs off the surface as if it were paved. Every tenth-of-an-inch of rain produces a rise in the river, and halfan-inch causes a considerable flood.

The course of the Tiber and its tributaries is almost entirely through impermeable strata: the Jura limestone in the mountains, and the clays, sandstones, and tufas of the tertiary and post-tertiary strata. Hence its floods suffer little diminution from the percolation of water through the soil.

4. NUMBER OF TRIBUTARIES.

The shrewd old tyrant Tiberius, when he attributed the floods of the Tiber to the multitude of its affluents,* shewed that he understood far more about the matter than any of our modern engineers, who fancy that these inundations are caused by a few petty obstructions in the bed of the stream, as if rivers in a state of nature never overflowed their banks. Pliny enumerates two and forty rivers which flowed into the Tiber below the confluence of the Chiana.† Most of these are mere brooks, and, to avoid confusion, are omitted in the map prefixed to this work, but, if they were represented, the map would present a complete network of streams.

It may seem at first as if all the rain which falls in the quadrilateral formed by two great tributaries, the hills in which they rise and the main stream must find its way into the river, with the exception of the part which disappears at once

* του τε ποταμου του Τιβέριδος πολλὰ τῆς πόλεως κατασχόντος, ὥστε πλευσθῆναι....ἐκεινος δὲ δὴ νομίσας ἐκ πολυπληθίας ναμάτων αυτὸ γεγονέναι, K.T..-Dio Cassius LVII. 3, Tib.

The river Tiber having inundated great part of the city, so that it was navigated by boats....he thinking that this was owing to the multitude of its tributaries, &c.

+ Infra Arietinum Glanim duobus et quadraginta fluviis auctus, præcipue autem Nare et Aniene, qui et ipse navigabilis, Latium includit a tergo. -Plin. III. q. II.

by evaporation and absorption. But this is not the case. In the steppes of Southern Russia, lying between the great affluents of the Don, small streams flowing into these affluents are rarely met with. Owing to the horizontal arrangement of the strata the water arising from the rain, after flowing for some distance, forms little ponds and morasses, and eventually either sinks into the earth or disappears by evaporation. On the other hand, the numerous little streams which intersect the basin of the Tiber convey the rain rapidly to the river before it has time to evaporate or sink into the earth. The result is the same as that produced artificially by thorough draining, which is said to have the effect of increasing the floods in winter, while it lessens the supply of water during the summer droughts.

It thus appears that the size and form of the basin of the Tiber, the large rainfall within its basin, the number of its tributaries, and the impermeability of the soil through which they flow, all conspire to increase the volume of its waters during seasons of heavy and continuous rain. To the increase of that volume the floods of the Tiber are owing, and not to the damming up of its waters by winds or artificial obstructions. When the rains are exceptionally heavy and extend at

once

over all minor basins drained by its tributaries, then we have a combination of circumstances producing those great inundations for which a remedy is being sought. But an efficient remedy will, I believe, be sought in vain.

In support of this conclusion let us examine the plans which have been proposed for preventing these inundations or lowering their height, so that they shall cease to be an inconvenience to the city.