1904-5.] THE GROWTH OF TRINIDAD. 145 
capacity and aid in rendering the outflow of water more regular and mor® 
gradual. From the main reservoir or that occupying the main valley 
there extend on every side into the hills accessory and subordinate reser- 
voirs each of comparatively small capacity but forming in the aggregate 
a very large water-holding area, each part of which is so connected together 
as to part with its contained water slowly through channels and fissures, 
sometimes so narrow as to admit scarcely the blade of a knife. 
Longitudinal Section.—Fig. 8 represents a section of the river-valley 
along its course and at right angles to the line of fault. NBS is the present 
level of the valley bottom, ME being the thrown-down upper portion. 
The area NBME is filled-up with alluvium, and is a part of the vast under- 
ground water reservoir before referred to. BE shows the extent of the 
downthrow along the line of fault. 
Similar phenomenon to those now described occur no doubt in other 
parts of the world, but with one exception, I have not met with any ac- 
count of them. They vary considerably from the ordinary form of water 
sources as described in the textbooks (e.g., Geikie, Geology, page 344, 
Lyell, Principles, Vol. 1, page 391). The exception I have referred to is 
contained in a paper on the water basin of Loch Derg in Ireland, by 
G. H. Kinahan (Geological Magazine, London, 1873, page 286). There 
is however, this difference in the two cases, that in ours one fault produces 
depressions in each of the valleys it successively crosses while in the case of 
Loch Derg, a number of depressions are produced in one valley by succes- 
sive faults. 
The course of a fault or dislocation is often obscured by the effects of 
denudation and meteoric action, and where it crosses valleys it is covered 
with alluvium, so that it is not always possible to follow it with any cer- 
tainty, but in other cases it is very evident by the broken ground, the 
escarpments or other natural features which mark its course often feet 
or yards wide. ‘This is why faults are so apt to become valleys, the 
broken-up material occurring along the course of the fault being more 
easily removed by denudation than the solid unfractured rock, and the 
water having easier access. These facts are of course, familiar to every 
geologist, but it is as well to mention them here. In certain parts of the 
island, e.g., Naparima, every valley (or at all events such as run east and 
west) is on the line of a fault. In the northern hills this is not so much 
the case as the principal valleys cross the lines of dislocation at right angles 
and only the subordinate (or east and west) valleys are situated in faults. 
Unless indeed, it should be hereafter found that any of the north and 
south valleys are in lines of fault; but of this there is no evidence; and it is 
