348 MR. L. LEWTON-BRAIN ON THE ANATOMY 
With so large a number of species included in this group, it is inevitable that they 
should cover a great range of habitat; some of those included might almost equally 
well be placed in Group ii.; others are characteristic of wet meadows, others of shady 
meadows, others of cornfields, others of dry pastures, and so on. On the other hand, 
none of them are found in very dry, and none in very wet, situations. 
Corresponding to this variety of habitat we get a fairly wide range of leaf-structure. 
Nevertheless, though there are no positive characters common to the group, there are 
many negative characters which will hold good throughout almost the whole of it. 
On the whole, then, the leaf-structure of the grasses of this group is not a highly 
specialized one; and this fact renders it of importance, because the leaf-structure of 
other groups can with advantage be compared with it. 
In form, the leaf-section varies from that of the two species of Poa, with the upper 
and lower surfaces perfectly flat and parallel, to that of Aira cespitosa, with the upper 
surface thrown into extraordinarily high ridges. The latter grass is, however, excep- 
tional in this respect, and only a few leaves have the perfectly smooth surfaces of Poa 
trivialis and P. pratensis; the great majority of the grasses of this group have leaves in 
which the upper surfaces possess a number of low, rounded ribs. In most cases these 
ribs are extremely low, indeed practically obsolete, but in a fair number the thickness 
of the leaf at the ribs rises to be double that at the grooves; this is the case, for instance, 
in Festuca duriuscula and F. rubra and in Koeleria cristata. As will be seen, however, 
from a glance at the figures, we get all intermediate stages between these leaves and 
those where the upper surface is practically smooth. The majority of the leaves 
are not markedly hairy; in a good many, however (Pl. 39. fig. 46, Pl. 37. fig. 24, 
&c.), we find a small number of short hairs or asperities, either confined to one side or 
occurring on both; moreover, in Festuca duriuscula (Pl. 37. fig. 17), and in one specimen 
of Koeleria cristata, we find that the leaf-surface (or surfaces) is covered with numerous 
long hairs; both these grasses occur in rather dry pastures, and the hairs may thus be 
a protection against excessive transpiration. Against this explanation must be set the 
fact that the leaf of Koeleria cristata (Pl. 36. fig. 15), from the dry sandy pastures of 
Gullane, is conspicuously less hairy than that taken from a plant cultivated under much 
moister conditions in the Royal Botanie Garden at Edinburgh. 
Again, in the majority of these leaves stomata are found both on the upper and on the 
lower surface ; however, in most of the leaves the stomata occur in greater abundance 
on the upper than on the lower side, and in Festuca duriuscula there are no stomata on 
the lower side of the leaf—another point in which this’ grass is seen to possess the most 
xerophytie leaf-structure of the group. 
Generally the cells of the lower epidermis are more strongly cutinized than those of 
the upper, but in the majority of cases the difference is not marked. Here, again, 
Festuca duriuscula is an exception ; so, also, are F. rubra and Koeleria cristata, as in 
their leaves the difference between upper and lower epidermis is great. 
The leaves are never specially broad. or narrow, neither exceptionally thick nor thin. 
The vascular bundles are never very numerous, nor are they ever very few; the smallest 
number of bundles are found in the upper leaves of Festuca rubra. 
