12 PASTURE STUDIES: SOME RESULTS. 
It is obviously impossible directly to compare the results for 
Areas 1-5 with those for Areas 6-8 as the natural factors are 
different for the two groups. For each of the groups, however, 
the soil and altitude factors may be regarded as relatively constant, 
and for Areas 6 and 7 the treatment of the land was so nearly 
similar that this also may be considered almost constant. The 
treatment of Areas 1-5, although mainly similar, was different in 
respects which may be of considerable importance, while the 
composition of the seeds-mixture varied considerably for each 
area. 
The aggregate results obtained by the P.c. fr. method are 
given in terms of sown species and those not sown, but where the 
seeds-mixture is widely different, the significance of these aggregate 
results becomes less. This is due to the fact that in no case 
probably would the soil be quite free from living plants or at least 
of seeds which might produce plants in the young pasture, when 
laid down to grass (3). In addition, no provision could be made 
against the recolonisation of the ground by indigenous plants 
whose seeds might be introduced by natural means. 
The term “ sown species” used here must therefore include 
all the plants of those species the seeds of which had been 
included in the seeds-mixture, whether their presence be the result 
of sowing the seed or not. 
The difficulty which occurs under this consideration is how- 
ever very much reduced by the fact that most of the areas have 
been examined in successive years. 
GRAMINEAE. 
(2) Sown. 
As will be seen from Table III., the tendency on all these 
areas is towards a decrease in the proportion of sown grasses from 
year to year. This is very clearly shown by the results for Area 2 
where there was a steady decrease from 10-5 per cent. at 54 years 
to 4-1 per cent. at rof} years. Assuming that this decrease con- 
tinued at the average rate of the last three years for the further 
period of three years (up till 13} years), the sown grasses would 
