180 The Rothamsted Grass Experiments. 



some, at any rate, of the supplied nitrogen remained within the 

 soil, and that some of this was available to the growing plants 

 during the succeeding seven years. Moreover, a glance at the 

 table will show that as much nitrogen was taken up as on the 

 plot where the ammonia-salts were still annually applied ; 

 but in reference to this point there is the significant fact that 

 the surface soil of plot 6 showed at the end of twenty years a 

 notably lower percentage of nitrogen than the corresponding 

 layer of plot 5, thus pointing to the soil itself as being the 

 source of nitrogen. The reader may here find it instructive to 

 refer back to the discussion of the results on plot 7, page 149. 

 Since the application of the mixed mineral manure, the 

 flora of plot 6 has become much more complex. Agrostis 

 vulgaris and Festuca ovina are less predominant, Holcus 

 lanatus has much increased, whilst Dactylis gloinerata, 

 A vena elatior, Avena pubescens, Lolium perenne, and Poa 

 pratensis are more favoured than on plot 5. Of leguminous 

 plants, the creeping and comparatively surface - rooting 

 Lathyrus pratensis has much increased, as has the legu- 

 minous herbage as a whole. With its increasing complexity, 

 the roots of the herbage would doubtless acquire possession 

 of a more extended range of soil and subsoil, and more 

 varied powers of underground food-collection would come 

 into play. Whilst, therefore, some part of the nitrogen of 

 the increased produce obtained on the substitution of the 

 mineral for the nitrogenous manuring would probably be 

 derived from the residue of the previous applications, it is 

 probable that the greater part would be due to increased 

 power of underground food collection, by virtue of which not 

 only the immediately preceding, but the earlier accumula- 

 tions, or what may be termed the normal stores of the soil 

 and subsoil, would be drawn upon. That such was really the 

 case is concluded from the fact of the reduction of the per- 

 centage of nitrogen in the surface soil of plot 6, as of plot 7, 

 where mineral manure alone had been applied for twenty 

 years, and where the very complex and highly leguminous 



