Feb. II. 1918 Effect of Season and Crop Growth on Soil Extract 325 



During the growing season selected for these experiments, from May 

 to September (the first season, however, being from June to October), 

 very little rain falls in Berkeley. To protect the soils from flooding by 

 any heavy showers that might occur, waterproof canvas covers have 

 been provided and have been put over the wire houses in a few 

 emergencies. During the season of heaviest rains these covers are kept 

 on, and the containers are never flooded or subjected to leaching by 

 rain. In this season between crops the soils are watered with dis- 

 tilled water at intervals, so that they are close to the optimum at all 

 times. It is, of course, true that this is not comparable to the seasonal 

 changes to which field soils are subjected. On the other hand, such 

 seasonal changes of rainfall and drouth are never comparable for any two 

 seasons or any two places. Many orchard soils in this State, by con- 

 stant cultivation and irrigation in the summer, retain a very regular 

 moisture supply in the soil. The moisture conditions of this experiment 

 are, in fact, very similar to those found in such cases. The orchard, in 

 addition, is subjected to occasional large excesses of water from rainfall. 



After filling the containers with the prepared soils, the excess portions 

 of sifted soil were stored in the set of tightly covered bins shown in 

 Plate 14, B. This soil was available for supplying the small portions 

 removed by sampling and also furnished a stock for additional studies. 



ANALYTICAL PROCEDURE 



In the season of 191 5 the analytical methods used were those outlined 

 in United States Department of Agriculture Bureau of Soils Bulletin 31 

 {50). With pure solutions containing only a single compound the de- 

 terminations obtained by these methods are very satisfactory. Before 

 the close of the season considerable doubt was felt in regard to certain 

 of the results obtained. This applied especially to calcium, potash, and 

 phosphate. Before entering upon the work of the season of 191 6 a series 

 of studies were outlined to test the accuracy of the methods which had 

 been in use. The preHminary work previously performed by the pro- 

 cedures in Bulletin 31 had all been carried out with solutions which con- 

 tained only one pure salt of the radical to be determined. Under these 

 conditions the results were extremely accurate. 



In the studies now undertaken three sets of solutions were prepared 

 to simulate as closely as possible the highest, average, and lowest con- 

 centrations ordinarily obtained in soil extracts. The salts used were 

 monocalcium phosphate, calcium nitrate, magnesium sulphate, potas- 

 sium chlorid, and ammonium chlorid. The solutions were made up with 

 distilled water which had stood in ordinary glass bottles and so con- 

 tained small amounts of sodium and silica, though not of course com- 

 parable to the amounts which occur in water extracts from soils. 



