476 University of California Publications in Agricultural Sciences [Vol.3 



despite the large correction that the consideration of the probable 

 error might introduce. 



The differences in the crop producing power of the soils are very 

 marked in the Diablo clay adobe, where the second crop, as well as 

 the first, shows evident variations in the ability to support a crop. 

 In the Altamont clay loam the second crop almost loses the variations 

 seen in the first crop from pot to pot. The samples of both types 

 seem to show one thing in common — the approach of the several sam- 

 ples toward a uniform ability to produce crops, as the soils are kept 

 for longer periods under the same conditions. The Hanford soils did 

 not show, with the several crops, the parallelism in the fertility from 

 crop to crop as did the Diablo and Altamont soils. Some soils pro- 

 duced good crops of grain and poorer crops of legumes, others did the 

 opposite. The low nitrogen content in this type seemed to be a limit- 

 ing factor. This would account for the variation between the grain 

 and the leguminous crops. Also, the presence, or absence of Bacillus 

 radicuola inoculation in this connection might greatly affect the total 

 crop produced. 



There does not seem to be much doubt but that the soils of the 

 several types compared in this way are not the same, though they are 

 in certain respects similar. 



The Place of Soil Classification. — With all these evidences that the 

 soils within the several types are not closely similar, though they are 

 classified the same by the Bureau of Soils, what conclusion is one to 

 reach as to the value of such a classification ? If it were true that there 

 were no appeal from the findings of such laboratory and greenhouse 

 determinations as these, and that these determinations were a final 

 proof of the fertility or infertility of a soil, obviously there would be 

 but one thing to do — discard all such field classifications as useless. 

 But the writer is one of a great many soilists who are not willing to 

 rely on laboratory or even greenhouse results for an absolute deter- 

 mination of fertility, and for the grouping together of soils into a 

 workable classification. Not enough is definitely known as to the mean- 

 ing of such findings, though there are certainly many valuable points 

 shown by laboratory analyses. 39 



As examples of the value of natural classifications we rnay con- 

 sider those of botany, zoology, or mineralogy. If available, a wholly 

 satisfactorj' classification of soils would be equally useful. The appre- 



39 Jordan, W. H., Measurements of Soil Fertility, Xe^v York Agr. Exp. Sta., 

 Geneva, Bull. 424, 1916. 



