24 PASTURE STUDIES: SOME RESULTS. 
on these areas to counterbalance its objectionable character as 
an arable land weed. 
Poa pratensis. 
On most of the pastures examined by Armstrong (1), this 
grass was absent, and only in one case was it found by him to 
reach = per cent. of the total herbage. 
Sutton (19) expresses the opinion that “it should be service- 
able in Wales,” but the present results give no confirmation to 
this view. 
Only on Area 2, and then only in very small quantity, was it 
sown, and its highest P.c. fr. on this area was 0-6, while it was not 
recorded on the area in three out of six examinations. It was not 
recorded on Areas 1 and 3, but it had a P.c. fr. of o-1 on Area 4 
at 134 years, and on Area 6 at 54 years. 
On all pastures examined by the writer its highest P.c. fr. 
was 3:2." 
Poa trivials. 
Armstrong’s results (I) show that the prevalence of this grass 
varies in some cases from month to month. He found its average 
range to be from o-8 to 5-5 per cent. of the total herbage on first- 
rate old pastures. 
Its seed was not sown on any of the present areas, but the 
grass was recorded in percentage frejuencies ranging from o-g to 
5:6 on Areas 1-4.2, It was most strongly represented on Area 3, 
but it was more strongly represented on Area 1 than on Area 2 at 
corresponding periods. On Area 1 it decreased from 4$ to 54 
years, but on Area 2 it increased almost regularly from 5% years 
to 104 years. The difference between Area 2 at ro4 years and 
Area 4 at 13$ years is very marked, and suggests that a rapid 
decrease would follow on Area 2. 
On these areas, however, the grass is fairly strongly indigen- 
ous, but on Areas 6-8, as shown by its absence from areas 7 and 
8, it is not indigenous under natural conditions. Yet it was 
1 Anglesey, an old pasture of good quality at 50 feet, 
2 This grass was not separated from P. annia on Area 3 at 84 years, 
