250 
F. W. Oliver. 
consists, as we said, of 5. raiiiosissiwa ; erect plants so closely 
grown as almost to touch one another. These of course are 
annuals and the zone is renewed each year. The ground they 
occupy is rich in humus and the evidence shews that at places the 
territory occupied has undergone considerable extensions. There 
is no tendency for the ground to become hummocky, nor—apart 
from Rhizoclonium —do other plants establish a footing. It seems 
evident that two different types of colonisation are in progress on 
the two sides of this marsh-spit and that these types should 
correspond to incipient phases of distinct associations. The further 
history of the Salicornia radicans-hum mocks is epitomised below. 
Fig. 16. Map shewing colonisation of a sand bank in the bed of a 
channel. The rhomboidal figures are the outlines of sand hummocks 
collected since 1903 by plants of Salicornia radiccins and .S', herbacca. The 
black transverse belts mark the 6'. radicaus , whilst the position of S. 
herbacca is marked by short vertical strokes. The hummocks accumulated 
by the latter are relatively small. The shading marks areas (hummocks) 
further colonised by Glycerin atari lima and Stneda maritima. These two plants 
do not appear to settle directly on the sand bank itself. The arrow points 
down stream. Seale *’ 0 . 
In view of so much evidence of advancing development it 
seemed desirable to construct a more detailed survey of this portion 
