STATEMENT BY M. CEZANNE. 223 



cubic metres, 300 to 350 cubic feet, that is almost the sixth part of 

 the Seine.* The Eure, the basin of which is in extent six times that 

 of the fountain of Vaucluse, gives no more water than this at low 

 water. 



" What a difference there is between the climate of Px'ovence and 

 that of Normandy ! In Provence the evaporation might carry off three 

 or four times as much water as the rain supplies ; the greater part of 

 the water-courses there are dried up in summer, and the plains of the 

 Camargue, notwithstanding inundations and irrigation, are covered 

 with saline inflorescence. If then, under this climate, the source of 

 the Vaucluse is so constant and so abundant, it is because its basin 

 being, as is the case with all sources, subterranean, the water which 

 has once penetrated the ground is there sheltered against evaporation." 



Now, turning to the consideration of forest lands, we find that, 

 besides the desiccating of swampy land by evaporating through the 

 stomates of the leaves the moisture taken up by the rootlets, another 

 effect of the growth of trees is to desiccate the superficial soil by 

 aiding the water which it receives to descend along its roots to a 

 lower level, whence it is less likely to be carried off quickly by 

 evaporation. 



" The roots of vegetables," says d'Hericourt, in a passage which is 

 quoted by Marsh, " perform the oflBce of draining in a manner 

 analogous to that artificially practised in parts of Holland and the 

 British Islands. The method consists in driving deeply down into 

 the soil several hundred stakes to the acre ; the water filters down 

 along the stakes, and in some cases as favourable results have been 

 obtained by this means as by hoi-izontal drains." 



And by Marsh it is remarked : — " It is an important observation 

 that the desiccation of trees by way of drainage, or external conduc- 

 tion by the roots, is greater in the artificial than in the natural wood 

 and hence that the surface of the ground in the former is not char- 

 acterised by that approach to the state of saturation which it so 

 generally manifests in the latter. In the spontaneous wood the 

 leaves, fruits, bark, branches, and dead trunks, by their decayed 

 material, and by the conversion of rock into loose earth, throuoh the 

 solvent power of the gases they develope in decomposition, cover the 

 ground with an easily penetrable matter extremely favourable to the 



* On the 12th August, 1858, fhe Seine was lower than had been observed for a 

 hundred and fifty-six years before. The gauging that day gave a delivery of 48 cubic 

 metres per second. The ordinary low water of the Seine is from 75 to 80 metres. 

 Belgrand, Annates des Fonts et Chaussiea. (1858 t. ii p. 222.) 



