The observations and reasonings of Belgrand and Valles, though their conclusions have not been accepted by many, are very important in one point of view. There writers insist much on the necessity of taking into account, in estimating the relations between precipitation and evaporation, the abstraction of water from the surface and surface-currents, by absorption and infiltration--an element unquestionably of great value, but hitherto much neglected by meteorological inquirers, who have very often reasoned as if the surface-earth were either impermeable to water or already saturated with it; whereas, in fact, it is a sponge, always imbibing humidity and always giving it off, not by evaporation only, but by infiltration and percolation. The remarkable historical notices of inundations in France in the Middle Ages collected by Champion [Footnote: Les Inondations en France depuis le VIe siecle jusqu'a nos jours, 6 vols, 8vo. Paris, 1858-64. See a very able review of this learned and important work by Prof. Messedaglia, read before the Academy of Agriculture at Verona in 1864.] are considered by many as furnishing proof, that when that country was much more generally covered with wood than it now is, destructive inundations of the French rivers were not less frequent than they are in modern days. But this evidence is subject to this among other objections: we know, it is true, that the forests of certain departments of France were anciently much more extensive than at the present day; but we know also that in many portions of that country the soil has been bared of its forests, and then, in consequence of the depopulation of great provinces, left to reclothe itself spontaneously with trees, many times during the historic period; and our acquaintance with the forest topography of ancient Gaul or of mediaeval France is neither sufficiently extensive nor sufficiently minute to permit us to say, with certainty, that the sources of this or that particular river were more or less sheltered by wood at any given time, ancient or mediaeval, than at present. [Footnote: Alfred Maury has, nevertheless, collected, in his erudite and able work, Les Forets de la Gaule et de l'ancienne France, Paris, 1867, an immense amount of statistical detail on the extent, the distribution, and the destruction of the forests of France, but it still remains true that we can very seldom pronounce on the forestal condition of the upper valley of a particular river at the time of a given inundation in the ancient or the mediaeval period.] I say the sources of the rivers, because the floods of great rivers are occasioned by heavy rains and snows which fall in the more elevated regions around the primal springs, and not by precipitation in the main valleys or on the plains bordering on the lower course.