cation became permanently closed by sand any considerable flow 

 of fresh water during flood would be checked. This would prevent 

 the fresh flood water from carrying coarse material into the lake, 

 the water having insufficient velocity to transport any but the finer 

 material, in this way the occurrence of fine material in the surface 

 soil and much sand in the lower layers would be explained. 



For purposes of comparison a few tests were made of four typical 

 classes of soil found on this area, ri-z. : 



(1) Heavy sxirface soil, This covered the larger proportion of 

 the total area. 



(2) Light surface soil. Found in small patches only. 



(3) Black subsoil at three metres depth. Fairly general at 

 three metres. 



(4) Sandy black subsoil at two to three metres. .Met with in 

 patches at two to three metres. 



No. 1 swelled very considerably when wetted, permeable. 



No. 2 proved to be very permeable. 



No. 3 very slightly permeable. The water which passed through 

 the soil was coloured. 



No. 4 scarcely at all permeable. Percolate very strongly coloured. 



Curves comparing the rates of percolation, under similar conditions 

 through soils Nos. 2 and 4 show a most remarkable difference (Fig. 1). 

 The mechanical analyses of these two soils are not very different, yet, 

 that from the surface is very permeable while the sandy soil from 

 two to three metres behaves like a clay. From the fact that the 

 percolate is coloured and deposits a flocculent precipitate of an organic 

 nature on standing, it would seem that the impermeability is due to 

 the presence of organic matter in a colloidal form. 



The percolation tests were made as follows. Filter papers, very 

 permeable, were fitted to Buchner funnels of 46 millimetres internal 

 diameter ; after wetting the papers, five grammes of the soil to be 

 examined was spread as evenly as possible on the paper. The 

 layer of soil was about two millimetres in thickness. The soil was 

 now wetted carefully and the funnels filled with distilled water. The 

 percolate was caught in graduated cylinders and the head of water 

 in the funnels was maintained constant l>y means of a Mariotte's 

 bottle suitably arranged. 



The same marked difference is also shown in the rate at which water 

 rises in these two soils by capillarity. A glass tube of 5 millimetres 

 bore was loosely plugged at one end and filled with the fine dry soil, 

 then gently tapped until the soil was compacted. The loosely plugged 

 end of the tube was then immersed in water. The water at once 

 commenced to rise in the soil as could be seen by the change in colour. 

 Daily or more frequent readings were taken of the height to which 

 the water had risen. These are shown in Figure 2. 



