2 5 1 
The Boiiche d' Erquy in 1907. 
of the marsh than could he obtained by a mere revision of the 
general map of 1904. For this purpose one of the old perpendiculars 
(forming the axis of the spit) from the original base-line was 
revived, and in relation to this the areas of interest were mapped 
by the “gridiron” method on a scale of ^g. By means of a series 
of long transects thrown across the perpendicular all strategic 
points can now be kept under close observation. 
To complete the history of the hummocks it is necessary to 
turn to one of the sand-banks in the bed of the principal channel 
in the Farm area (known as “Stream B”). 1 In September, 1903, 
when this sand-bank first came-under observation, it bore nothing 
but a thin scattering of small isolated plants of Snlicomia radicans 
and herbacen. Since that date colonisation has advanced rapidly 
(see Fig. 16). The 5. radicans plants, being perennial, have gone 
on collecting sand from the channel as the sand-laden tide rushed 
over them, till they have grown into hummocks eight or ten inches 
in height, and three to five feet in length. These hummocks, which 
tend to run together into hummock-systems (as at the left-hand 
bottom corner of Fig. 16), have undergone secondary colonisation 
by Obioue, Glycerin and Suada. The more advanced amongst these 
covered hummocks are to all intents and purposes identical with the 
Grass-Suaeda hillocks that form so considerable a feature upon the 
more stable portions of the Farm area. Inspection of the adjacent 
parts of the covered marsh shews that much of this tract has 
developed in the same way. The sand-bank, as it was in April, 
1907, is represented in the grid map reproduced in the accompanying 
Fig. 16. The grass-covered system of hummocks on the left is 
quite ready for incorporation in the settled marsh so soon as the 
stream abandons the sand-bank in the progress of its meandering. 
The rapid growth revealed by this sand-bank—essentially a 
concise epitome of what is happening on the spit—depends doubt¬ 
less on the favourable position which it occupies in the bed of a 
channel up which the sand-laden tide rushes with considerable 
velocity. The extension of the 6\ radicans in belts at right-angles 
to the direction of flow is a striking feature, the physiological signi¬ 
ficance of which stands in need of elucidation. Of its efficiency as 
a sand collector there can be no question. The whole mechanism, 
including the transverse position of the belt, closely resembles the 
collecting of sand on a sand-dune by Psamma, if water-carriage be 
substituted for wind-transport. The map (Fig. 16) brings out 
1 This channel forms the axis of the complex system of water¬ 
courses depicted on the left of Fig. 15. 
