358 Journal of Agricultural Research voi. xii.no.6 



The figures obtained in the preceding season for the mean and maxi- 

 mum variation between duplicates will be used in order to find w^hat will 

 constitute a significant difference in yield. By taking the mean varia- 

 tion as the standard, soils 5 A, 8A, lA, 6A, and 2A are to be considered 

 equally productive, while iiA and 14A are slightly lower. If, however, 

 the consen/ative figure of the maximum variation between duplicates is 

 used, it will be seen that No. iiA and 14A will also fall in the group of 

 the highest yield. These relations are presented diagrammatically in 

 figure 7. 



It will be seen that six soils were significantly lower than the group of 

 highest productivity. Of those No. loA, 4A, and 7 A were practically 

 equal, while 3A, 1 2 A, and 9A were distinctly low in yield. 



The results of the water extractions are again presented in the form of 

 graphs. In figures 8 to 20 the plot marked "A" is in each case the planted 

 portion, while the uncropped duplicate is called "B." Some extremely 

 striking differences are exhibited by these graphs. 



In the planted soils the water-soluble nutrients at the beginning of the 

 season, either remained on practically the same level or increased slightly 

 for the first four weeks. Then, without exception, the nitrogen com- 

 menced to decrease rapidly and was followed in a smaller degree by the 

 calcium and potash, and very slightly by the magnesium. The con- 

 trast shown by the unplanted soil was equally uniform in nature, though 

 variable in the extent of the effect. In almost all the soils the effect of 

 the cultivation was to cause a considerable liberation of soluble nutrients 

 and, though there was later shown to be a depression from this high 

 figure, yet the general range of all the nutrients except the phosphates 

 continued to be higher in the uncropped soil. 



The soils which did not show this stimulation and liberation of nutri- 

 ents were No. 3B, 9B, and 12B. These were the least productive soils 

 of the group and also had the lowest range of soluble nutrients. 



It is extremely significant that this same period of high soluble nutri- 

 ents in the uncropped soils corresponds to the period of lowered nutrients 

 in the cropped duplicates. It was also the period, as may be observed 

 in figure 21, in which the plants were making their most rapid growth. 



The one compound which did not exhibit this liberation of excess 

 nutrients in the uncropped soil was the phosphate. There were large 

 differences between the amounts present in various soils, but for any one 

 soil the amount was practically constant in the cropped and uncropped 

 plot. It was the only soil nutrient studied which behaved in this manner. 



All the above differences between the planted and unplanted soils 

 which have been noted are greater than can be accounted for by the 

 probable factor of error which has been determined in an earlier portion 

 of the study. The maximum figure for any nutrient in the concentration 

 found is 10 per cent, and that applies only to potassium and calcium. 

 On applying this correction plus or minus to the periods of greatest de- 



