1366 
MESSRS. J. B. LAWES, J. H. GILBERT, AND M. T. MASTERS, 
separation-year; Anthnscus sylvestris has gradually and considerably increased; 
Conopodium denudatum has very much diminished; and the several species of 
Ranunculus have become practically absent from the samples. 
Comparing the actual yield per acre on plot 14 with that without manure, the table 
shows that on the average, and almost in every individual instance, there was a con¬ 
siderable excess of every grass occurring in more than 1 per cent, excepting Festuca 
ovina and Agrostis vulgaris , which are the two most prominent grasses of the locality 
under poor conditions of the soil. Those which have yielded the greatest average 
increase compared with the unmanured produce are most prominently and uniformly 
Poa trivialis and Bromus mollis, next in order coming Dactylis glomerata, Lolium 
perenne, Alopecurus pratensis, and, less prominently, TIolcus lanatus, Arena jlavescens, 
Poa pratensis, and Arena elatior. 
Of Leguminosse there was, with the one exception of Lathyrus in two of the years, 
a large deficiency of each species on plot 14, compared with plot 3 without manure. 
Of Miscellanese there was, on plot 14 compared with plot 3, an excess of Rumex 
Acetosa, Anthriseus sylvestris, and, generally, of Conopodium denudatum ; but there 
was a deficiency of the species of Ranunculus , and a very marked deficiency of a large 
number which occurred in too small quantity on plot 14 to be included in the list of 
those ever yielding 1 per cent, to the mixed herbage. 
Compared with plot 9 (with the ammonia-salts), Poa trivialis said Bromus mollis are 
again the chief grasses in excess, and both are very largely so. In a much less degree, 
Lolium perenne and Alopecurus pratensis are in excess on plot 14; whilst Festuca 
ovina, Poa pratensis, and Agrostis vulgaris, are all largely, Arena elatior considerably, 
and Holcus lanatus, Dactylis glomerata, and some others in some degree, in defect. 
Small as is the actual amount of leguminous herbage on plot 14, still about half of it 
is in excess of that on plot 9, and the excess is due, as already intimated, to Lathyrus 
pratensis. 
Of miscellaneous species, the yield per acre of Rumex Acetosa, and of Conopodium 
denudatum, averaged considerably less on plot 14 than on plot 9 ; whilst the whole of 
the yield of Anthriseus sylvestris with the nitrate was in excess of that found in the 
samples from plot 9. 
The foregoing details show—and the table should be further examined on the point 
—that there was a very striking difference in the growth, and in the botanical com¬ 
position of the herbage, accordingly as the nitrogen (in conjunction with minerals) was 
applied as ammonia-salts or as nitrate of soda. With the nitrate there was on the 
average a notably greater quantity of gramineous herbage produced, somewhat more 
leguminous, but less miscellaneous herbage. With the ammonia-salts a greater number 
contributed a fair proportion to the herbage, and the most prominent grasses were 
Poa pratensis, Festuca ovina, Agrostis vulgaris. Arena elatior, and Holcus lanatus. 
With the nitrate of soda only one of these was in any special prominence; the most 
prominent were, instead of Poa pratensis, P. trivialis and Bromus mollis ; and the 
