— 20 - 



dense area, are improved thereby; inasmuch, as a certain portion of 

 water can be actually squeezed, into the upper part of the compressed 

 layer by intense pressure, and the seeds can benefit inmediately by this 

 moisture. So that subsoil rolling is beneficial only just before the sowing 

 of winter corn. As regards capillary ra-'sing of water from the deep layers, 

 the less said about it the better. 



On the other hand the constant movement of water downwards, 

 both in the upper and lower, even in very deep layers, cannot be de- 

 nied. In my treatise of the circulation of water in the soil -). I gave nu- 

 merous examples of the downward percolation of water at big depths 

 from 95 to 105 cm. (3 cases) and an insigneficant depths -from 35 to 

 45 cm. (2 cases), at a speed of 15 to 20 cm. per month'-*) on old 

 ploughed fields and twice as slowly on unploughed soil. 



All the data at my command regarding moisture on the soil of the Odessa 

 Experimental field, point only to one conclusion viz: that water percolating 

 beyond a depth of 40 to 50 c. m. does not return to the surface except 

 by the way of roots; all the water not seized by the roots goes down 

 into the deep layers, moving at the rate of about 7 feet yearly. In order 

 to verify this theory a pit was dug 1040 cm. deep on unploughed land; 

 that is on soil covered with vegetation. And as this pit lay in a hollow, 

 where rain water collected but had no outlet, there was a certain 

 surplus of water every year, which not being Imbibed by the roots, per- 

 colated into deeper layers, as stated above at a speed of about 2 metres 

 per year (table II). At 260 to 330 cm. remains of the previous years 

 1903, water were fonnd and at 460 to 530 cm. the surplus water which 

 had percolated thither in the course of 2 years (from 1902) previously. 

 If it be taken into consideration that the extreme limit of the Super- 

 ficial humid layer was fonnd to be at a depth of 130 cm., then the 

 foremost or leading part of the previous years (1903) humid layer should 

 have been abont 330 cm. deep but was fonnd at 330 cm. The fore- 

 most part of the 1902 layer should have been about 530 cm. deep and 

 was found at that depth. In both cases the remains of each year's wa- 

 ter have a distance of about 2 metres between them. 



Therefore there is no doupt about the process of water transference 

 from above downwards and its significant preponderence over the pro- 

 cess of the supposed capillary raising of soil water to the surface. The 

 formation of ground water goes on, not only at the expence of water 

 oozing into the depths from ponds, hollows and snow water in planta- 



■•) Circulation of water pp. 30 and 81. 

 "'■■''■•) Ditto page 32. 



