340 Journal of Agricultural Research voi. xii. No. 6 



taken with the iS-inch cheese trier previously mentioned. From 9 to 

 12 triers of soil were required for a sample of approximately 600 gm. 

 The holes left by removing these samples were filled with reserve soil 

 from the bins. 



It was felt that this procedure might be open to some objections, 

 owing to the fact that the soil in the bins was partially air-dried. In 

 any one season 10 sampling periods is the maximum number, and the 

 amount removed would not exceed 6 kgm. This amount was less than 

 0.8 per cent of the total soil present. The error involved was therefore 

 smaller than it was possible to detect by any available methods. 



In order to minimize this influence, the containers were always sampled 

 by a definite system, so that no two samples were taken from the same 

 spot twice in the same season. The process of sampling was performed 

 with great care. There was no breakage of plants and no evidence of 

 injury from the small roots cut by the cheese trier. The growth obtained 

 has been extremely satisfactory, both in height and vigor. The plants 

 were spaced at the usual field distance, 6 inches apart each way. In 

 1 91 5 one seed was planted in a place, and a small number of extra seeds 

 were started simultaneously in the same soil. These extra plants were 

 then transplanted into any spaces where the seed failed to sprout. An 

 even stand was obtained in all except two duplicate containers, soils 

 3 and 5. The following season the writers were able to obtain a per- 

 fectly uniform growth by planting three or four seed in each place, and 

 then thinning to one plant of average size. The seed used was furnished 

 by the Division of Agronomy and was a pure strain of Beldi barley. 

 Before planting, it was treated with formalin to prevent smut; seed of 

 uniform size were selected. 



A slight difficulty has been experienced toward the close of each 

 season from the tendency exhibited by the crops on several soils to lodge. 

 This has been corrected by a system of supporting wires. 



In 1 91 5 the crop matured, and was harvested from October 21 to 

 November 6. This was a later date than commercial grain crops are 

 grown in California, but it is believed that the results were strictly com- 

 parable among themselves. The crop yield is given in Table VIII and 

 is expressed as total air-dry crop and grain in grams. The grain is also 

 calculated as pounds and bushels per acre. It will be observed that in 

 all the soils except No. 3 and 5 the agreement of the duplicates was 

 excellent. These two soils, it will be remembered, were those in which 

 the growth was least uniform. 



