108 Observations and Experiments on some English Pasture*. 
tant elements in the pasture not to be found in the enclosure, 
on the whole, these plots fairly represented the general pasture. 
The results of this inspection, and the analysis of the herbage 
of the enclosed plots, convince me that small, transplanted turfs 
do not supply data to enable one to determine the composition 
of the pastures from which they are taken. Indeed, in the 
cases recorded in this paper the information they give is 
entirely misleading. Dr. Fream's data are accurate. He has, 
of course, given faithfully and accurately the different plants 
which grew on the turfs at Downton, and the various pro- 
portions of these plants. But these data have not been suf- 
ficient for the generalisations based on them as to the pastures 
from which they were obtained. The difficulty in obtaining a 
turf like those planted at Downton, that would represent the 
pasture, is to find within the space of two square feet specimens 
of the various plants, and in their relative proportions, which 
make up the pasture. In the middle of April, 1889, I visited 
some of the pastures in Romney Marsh. In a good field, which 
was well stocked, and, consequently, closely eaten down, I had 
two turfs the same size as those of Dr. Fream, growing quite 
close to each other, dug out, and transplanted where they could 
be observed. The farmer, and his foreman, who was so good as 
to dig up the turfs, considered each as an equally good specimen 
of the pasture. When the herbage of the turfs was cut in the 
autumn, and the elements separated, it was found that it con- 
sisted entirely of grasses, though the composition of the two 
turfs was very different. The following is the result of the 
analvsis : — 
Turf 1. Tttkf 2. 
Tci cent. Per cent. 
Rye-grass, Loli/im perenne . . 51 : Cocksfoot, Dactylis glomerata . 87 
Yellow oat-grass, Aoena Jla- ) lye-grass, Loliian perenne . 8 
vescens 32 i Yellow oat-grass, Arena jia- 
Dogstail, Cynosurus cristutus . 17 j vescens 4 
| Dogstail, Cynosurw cristatus" . 1 
It is obvious that generalisations based on these two 
neighbouring and representative turfs would lead to very dif- 
ferent conclusions. It is only by such an accident as would have 
happened in relation to the field in Romney Marsh if I had 
brought away only one of the turfs, that I can understand such 
a result as that shown by the Downton turf from North Curry 
(p. 100). The farmer at Romney Marsh would have considered 
either turf which he happened to fix on as representative, and 
sent it on to me for observation, had it been left to him to select 
the turf. At North Curry there was a considerable quantity of 
yaiTow in the field, but of course it was not so equally distributed 
