THE COLLOIDAL PROPERTIES OF SOIL 165 



and is then dissolved as ferrous bicarbonate. At the lower 

 level the COg escapes and the iron carbonate is precipitated 

 and then oxidised. 



These explanations seem very simple but they present 

 several difficulties. In the first place, Morison and Sothers 

 could find no evidence that peat or humic acid can reduce 

 ferric oxide, the first stage in the process, according to Mayer. 

 Ferric chloride is easily reduced : so easily, indeed, that ferrous 

 iron is produced during the process of testing a mixture of 

 ferric oxide and peat if one begins by extracting with HCl ; 

 but ferric oxide is not. Secondly, it is difficult to understand 

 why the deposition should be so local and so sharply defined. 



Recent workers, therefore (Munst,^ Ramann,^ Morison, and 

 Sothers (204), regard the whole process as a formation first 

 of a " sol " and then of a " gel," and Morison and Sothers 

 suggest the following as the most probable course of events. 



It is well known that "sols" change to "gels " in presence 

 of small quantities of electrolytes, and conversely " gels " often 

 change to "sols" when electrolytes are removed. In normal 

 soils the conditions are favourable to gel formation, but when 

 in these particular soils the upper layer of sand becomes de- 

 nuded of its soluble material by the persistent washing of rain 

 water, the conditions become favourable for the formation of 

 sols of ferric hydroxide and of humus — or ferric humate, if 

 one likes to put it that way. Morison and Sothers actually 

 obtained such sols^ by persistent washing of ferric-humus 

 gels. 



As the sol is washed down some is deposited on the 

 bleached sand, but the bulk of it passes to the permanent 

 water level where it remains and accumulates, diffusion being 

 practically non-existent. During the dry months a certain 

 amount of desiccation takes place, involving a deposition of 

 the sol as a gel : there is also a certain amount of transformation 



1 Miinst, Bied. Zentr. Agrik. Chem., 1912, 41, 3-10. 



2 Ramann, Bodetikunde, Berlin, igii, p. 204. See also H. Stremme, Kolloid 

 Zeitsch., 1917, 20, 161-168. 



' As might be expected these did not give the ordinary iron reactions. 



