Further Observations on Heath Association. 127 
Prostrate shoots of U. nanus were tied up in a vertical position 
in considerable numbers, and these showed a marked tendency, after 
some months, to die away at their tips. Several of these exhibited 
a downward curvature of the tied-up shoot, the tips in these cases 
remaining healthy. A similar dying hack of the tip was also observed 
in tied-up Erica cinerea. 
Since Ulex europceus is practically confined to the valleys, 
where it normally finds sufficient shelter to develop in the usual way, 
good examples of the prostrate type of this species are almost absent 
except on the land-slides, where they are in part due to rabbit¬ 
trimming (cf. below). There can be no doubt that there is a much 
greater tendency towards the production of prostrate branches in 
U. nanus than in U. europceus. Indeed careful investigation has 
shown that all specimens of the former, even those with a tall 
upright stem found in the valleys (3, p. 151), possess a creeping 
branch-system whose remains can be found hidden among the humus 
at the base of the erect shoot. These plants have then in all cases 
exhibited initial prostrate growth. 1 
In U. europceus , on the other hand, the tendency is towards the 
upright habit, and consequently the tall-growing specimens in the 
valleys do not possess branches prostrate on the ground, but realise 
the cushion-habit, in the more exposed situations, on branches which 
have over-topped the surrounding vegetation. Where plants of U. 
nanus have produced erect shoots, which become exposed, they 
likewise assume this densely branched character, either forming 
typical cushions or exhibiting a more or less corymbose arrangement 
of the upper branches, features which may perhaps correspond to 
differing conditions of exposure. The cushion-habit of such branches 
(PI. I, Phot. 5) is due to the regular dying away of the exposed tips 
(in U. europceus this amounts to inch in each annual increment) 
and the resulting stimulation to growth of the axillary buds immedi¬ 
ately below. At the same time the spinous branches are shorter 
and all the parts are smaller (cf. a and b in PI. I Phot. 7). In non- 
exposed branches (erect non-exposed type, cf. below) irregular dying 
back occurs, but the axillary buds sprouting below develop unequally 
and their average annual rate of growth is much greater; also the 
main axes of the branches tend to persist and often grow for more 
than one season. In other words the growth of exposed shoots is 
almost exclusively sympodial, whereas sheltered specimens are in 
1 It was previously thought that procumbent branches were generally 
absent in the upright form (cf. 3, p. 15‘2). 
