342 H. I. JENSEN. 



the arid ones. The reason for this is not clear, and the 

 accumulation of further data may reverse the result. 



Our own experience in the Department of Agriculture of 

 New South Wales is that leached laterite soils like those of 

 The Dorrigo, the Robertson Range and the Macpherson 

 Range, even when derived from highly manganiferous form- 

 ations, are low in manganese and high in humus. Coastal 

 marsh soils, highly acid peat soils, are quite free from 

 manganese. Manganese is, on the other hand, high in soil 

 from undrained depressions having a local catchment and 

 no outlet except in flood time. 



In alkaline soils an accumulation of manganese might be 

 expected when the alkalinity is due to carbonate of soda. 

 This has been found to be the case, and is due to the fact 

 that manganese enters largely into the composition of 

 colloidal clay (see Hilgaard). 



The manganese estimated during the course of my work 

 was only a portion of the total present, for the portion 

 which was carried down with the iron was not separated 

 and added on. The figures given therefore only repre- 

 sent from one-half to at most two-thirds of the manganese 

 present. Olearly these soils are high in manganese, a fact 

 suggesting that they have been accumulated in an un- 

 drained basin. The salt and carbonate of soda present are 

 evidences in the same direction. The remarkable thing is 

 that the faintly acid gilgai soils Nos. 5 and 15 are most 

 highly manganiferous, while the strongly alkaline soils 

 Nos. 4 and 11 are least so. This paradox is rendered less 

 startling when it is stated that the chemical work proved 

 the alkalinity due mainly to carbonate of lime, the soda 

 present being never a very excessive amount, though quite 

 sufficient to be injurious. 



It appears then that while the lime which creeps into 

 the hummocks by capillarity is precipitated there and 



