164 EXPERIMENTS UPON GRASS LAND 



soda, still grows a very large crop, averaging 46 cwt. over the 

 whole period. The vegetation resembles that of Plot 14, but 

 there is much less False Oat than on 14 3 per cent, instead of 

 40 per cent. and it is more varied, there being about four times 

 as much leguminous herbage, among which the Meadow 

 Vetchling predominates. This plot probably marks the limit 

 of the amount of nitrate of soda which it would be profitable 

 to apply in ordinary farming, since the second 275 Ib. per acre 

 of nitrate of soda on Plot 14 has only produced an average 

 increase of 11 cwt. of hay. 



Reviewing the whole of the evidence, nitrate of soda is 

 distinctly a better manure for hay on the Rothamsted soil 

 than are ammonium- salts, producing more grass and that of 

 a better quality. 



On Plot 10 the potash is omitted from the mineral manure, 

 though the other minerals and the nitrogen are the same as on 

 Plot 9. The result of the omission of the potash is a consider- 

 able decline in yield, which has become more accentuated as 

 the experiments have progressed and the original stock of 

 potash in the soil has been reduced. The herbage consists 

 even more wholly of grass than does that of Plot 9, and the 

 development of flower and seed is distinctly later. 



Plot 4-2 receives the same ammonium -salts, supplying 86 

 Ib. of nitrogen, and superphosphate only, so that it compares 

 with Plot 9, except for the entire absence of alkaline salts. 

 The lack of potash shows itself in a great reduction of crop, the 

 average over the whole period having been only 33 cwt. against 

 54 cwt. on Plot 9. It is thus much below Plot 10, also with- 

 out potash, but which receives magnesia and soda. The herbage 

 on this plot again consists almost wholly of grasses, which have 

 a very dark green colour and are late to mature. The dwarf- 

 growing and shallow-rooted grasses predominate ; Sheep's 

 Fescue constitutes more than one-half, and, with Sweet Vernal 

 and Bent Grass, as much as 03 per cent, of the whole herbage. 



The characteristic appearance of the herbage is well seen in 

 the photograph, Fig. 27, of turf taken from this plot in 1903. 



