155 



badly. In the Cape Division, too, it will be seen that the analyses by 

 Method II. give much higher results on the whole than those by Method I, 

 Clanwilliam and Knysna are Divisions where the Table Mountain sand- 

 stone series predominates; and yet here too the second method of extrac- 

 tion took considerably more lime and potash from the samples subjected 

 to this treatment than could have been done by crops. In the sandy 

 soils of Eerste River in the Stellenbosch Division this difference is also 

 visible. 



In the Hanover soils may be noted the large quantities of lime ex- 

 tracted by the second method from soils belonging to the Karroo system. 

 No analyses of any soils from this area have been made by the first 

 method, but where that method was applied to soils of more or less 

 the same nature, for instance those from the neighbouring districts of 

 Aliwal North, Albert, Graaff-Reinet, Richmond, and Colesberg, more 

 limited proportions of lime were extracted. In the Division of Steyns- 

 burg, where a number of soils were analysed by both these methods, a 

 similar disproportion with regard to lime will be noticed. 



Something of the same kind may be observed in respect to the soils 

 of the Robertson Division, where, however, the geological formation be- 

 longs to a different system. The Robertson soils analysed by Method I. 

 showed a distinctly lower average of lime. 



The amounts of the phosphoric oxide extracted by Method II. do not 

 differ materially from the proportions extracted in cases where Methods 

 I., III. and IV. were applied : in these latter cases, it will be noted, the 

 phosphoric oxide was not determined in the hydrochloric acid extract, 

 but was extracted by a special process. 



In one case (a granite soil from Kuiken Vallei, Stellenbosch Divi- 

 sion) the extraction process was varied, boiling Hydrochloric acid being 

 employed: apparently, however, no larger quantities of plant food con- 

 stituente are thus removed from the soil than by Methods II. and III., 

 but it is difficult to avoid coming to the conclusion that the method 

 adlopted as a standard, fitting in as it does so well with the opinions of 

 practical men, is, on the whole, a much safer guide. 



Of course, it is evident that, under certain circumstances, Methods 

 II., III. or IV. may not extract from the soil any more of a certain plant 

 food constituent than Method I.; but that may be because Method I. 

 extracts all that there is to extract. This may be expressed by saying 

 that in such a case probably all of that plant food constituent in the 

 soil falls easily within the limits of what has been termed grade II.* or 

 the reserve stock. 



The poor Witteberg soils of Albany, which yielded so small amounts 

 of phosphoric oxide to Hydrochloric acid, naturally could not be expected 

 to show to advantage by the citric acid method, and the results thus 

 obtained are of course exceedingly low, but one of the three soils, which 

 gave a fair proportion of phosphoric oxide on extraction by Method III., 

 yielded '026 by Method V., in other words the result was> well over Dyer's 

 proposed minimum limit of '010 per cent.f In one of the soils collected 

 at Elsenburg, Stellenbosch Division both potash and phosphoric oxide 

 showed poorly only '024 per cent in each case by Method I. : in this 

 instance Method V. yielded '017 per cent of potash, but only '0036 per 

 cent, of phosphoric oxide, while extraction with water gave only "0013 

 per cent, of the latter. The soil was, according to both methods, clearly 

 lacking in both these two elements of plant food. The three soils from 

 the Herbert Division extracted by Methods I. and V. afford further in- 

 stance of mutual confirmation. On the basis of Method I., soil No. 1 



* See page 9. t See page 17. 



