222 The Rofhamsted Grass Experiments. 



of the different seasons ; and, as far as may be, to ascertain 

 what are the special characters of growth above ground or 

 under ground, normal or induced, by virtue of which the 

 various species have dominated, or have been dominated over, 

 in the struggle which has ensued." At the outset it was 

 noticed that those manures which are most effective with 

 cereals grown on arable land, were also most active in 

 increasing the quantity of grass amongst the herbage, and 

 that the manures which are most beneficial to beans or clover 

 produced the greatest proportion of leguminous herbage. 

 Thus, the highest gramineous produce resulted from a highly 

 nitrogenous manure, such as ammonia-salts or nitrate of 

 soda, with alkaline-salts, particularly potash ; but side by 

 side with the increase in the total gramineous herbage, there 

 was a decrease in the actual number of species of grass. 

 On the other hand, the highest percentage of leguminous 

 produce was the result of a mixed mineral manure with 

 potash. The percentage (by weight) results on the following 

 plots illustrate these points : 



Plot 7 was the most favourably manured for leguminous 

 produce, it received mixed mineral manure alone, including 

 potash ; plots 3 and 12 were the two unmanured ones ; plot 

 11 was the most favourably manured for gramineous produce, 

 it received SOOlb. ammonia-salts per acre, with mixed mineral 

 manure, including potash. 



Special observations and complete botanical separations 

 made at intervals of five years to determine the influence of 

 seasonal variations show that " a given quantity of the pro- 



