1770 FORAGE CROPS, MEADOWS AND PASTURES 



le}'' than it otherwise would, but there is always some risk of the sown 

 plant interfering with the development of the definitely lasting indigenous 

 species. Good examples are Poa trivialis and Cynostirus cristatiis ; a 

 knowledge of general conditions would often suggest excluding the former 

 from a mixture, while the amoimt of seed of either that might be advan- 

 tageously used needs further local investigation. 



The desirability or otherwise of sowing the commercial seed of 

 locall}- exotic species, or of but slightly secondary species, is easily deter- 

 mined. The commercial seeds lead to good results or they do not. If 

 the commercial seed produces plants that are found to succeed it is however 

 nearly always necessary to sow liberally ; this is particularly true of Dac- 

 tylis glomerata, Festuca elatior, Cichorinm Intyhus, Phleiim pratense and 

 A rrhenathenim avenaceum. 



It is difficult to account for the spontaneous appearance of primary 

 and secondary species (and of some locall}^ exotic species, as Holcus lanatus, 

 Bromtis mollis et spp. and Phleum pratense,) on land long under the rotation 

 when put down to grass, especially when more or less isolated from natural 

 or semi-natural grasslands. The available evidence, however, suggests 

 that : 



a) Many species remain on the land as arable weeds on the soils 

 that suit them, as Poa trivialis, Agrostis vulgaris, Festuca rubra and Ra- 

 nunculus repens. 



h) The seeds of many species are probably introduced by the wind, 

 as Holcus lanatus, Anthoxanthum odor alum. 



c) The seeds of many species are certainh^ introduced as impurities 

 (useful and otherwise) in the sown seeds, as for example Phleum pratense 

 (in Trijolium hybridum), Plantago lanceolata (in Tri folium spp.) Holcus 

 lanatus, and Bromus mollis et spp. (in Lolium spp.) 



d) There seems little doubt that the seeds of a great number of 

 species are capable of lying dormant for long periods in the soil ; in parti- 

 cular this m.ay be true of Trifolium repens and T. minus, Cynosurus cris- 

 tatus, Poa spp, Phleum pratense and Festuca ovina. 



The data brought forward in this paper would seem further to justif}'' 

 the following broad generalisations with regard to both experimental 

 work on grassland and the whole problem and economics of putting 

 land down to grass. 



i) Experimental plots dealing with seed mixtures should be large 

 (at least half an acre) and square in order to give a considerable central 

 zone. The hay should be cut as early as possible to avoid seeding and the 

 carriage of seed from plot to plot. No series of plots can, henceforward, 

 be regarded as complete without a control plot, which control should not 

 be seeded (with grasses and clovers) biit left to the indigenous species to 

 colonise. The control plot should, of course, be subjected to the same 

 cultivations, receive the same manures, and grow the same nurse as the 

 seeded plots. 



2) Undoubtedly when putting land down to long duration grass 

 as much or more can be done b}^ making the habitat as suitable as possible 



