154 EXPERIMENTS UPON GRASS LAND 



becomes important to ascertain the natm-e of the plants 

 comprising the herbage produced by a given manm^e, as well 

 as to determine its amount ; from time to time therefore at 

 Rothamsted a carefully selected fraction of the herbage from 

 each plot has been separated into its constituent species, the 

 relative proportions of which are determined by weighing. As 

 this complete separation involves a great amount of work, a 

 partial separation only is made every year, in which case the 

 herbage is separated into three groups — the grasses, the 

 leguminous plants, and the miscellaneous species respectively. 

 Table LIX. shows the results of these partial separations 

 as averages for the whole period of forty-seven years, and 

 for the single year 1902. Summaries of the five comjDlete 

 separations made in 1862, 1867, 1872, 1877, and 1903 are 

 given in Table LXII. (see p. 173). 



I. The Unmanured Plots. 



Two of the plots have remained without manm'e during 

 the whole of the experiment. They are situated near the 

 extremities of the field, and show a slight but constant 

 difference in crop. Taking the average of the whole period, 

 these unmanm-ed plots have produced rather more than a ton 

 of hay per acre per annum. If we compare the successive 

 ten-year returns, there is no sign of approaching exhaustion or 

 great falling-off" in crop from year to year. The impoverishment 

 of these unmanured plots is more to be seen in the character 

 of the herbage than in the gross weight of produce. Weeds of 

 all descriptions occupy the land, and the relative proportion 

 they bear to the grasses and clovers has increased from year to 

 year. A fair proportion of clovers, both red and white, is 

 found on these plots, l:)ut the weeds, which amount to 26 per 

 cent, taking the average over the whole period, have of late 

 years constituted nearly one-half of the herbage. Tlie most 

 prominent species among the grasses are the Quaking Grass, so 

 generally taken as a sign of poor land, which constituted 20 

 percent, of the whole herbage in 1903, and Sheej^'s Fescue; 



