166 J. STANLEY GARDINER. 
of nullipores as a factor in their formation may be practically neglected. The importance 
of these organisms, when present, in welding up the different coral colonies cannot be 
over-estimated. In effect they act as the ties which keep the whole together, so that it 
ultimately forms the apparently solid reef. In the case of Mahlos much of the so-called 
reef-tlat round islands may have been part of the original raised limestone, but it is certain 
that no regular flat is now being formed without the cooperation of nullipores. The inner 
reefs are mere agglomerations, formed by the growths of various corals, their interstices in 
some measure filled in by Halimeda remains, coral and shell fragments and sand. The 
corals are bored into by sponges, algae, worms, ete., and killed at their basal ends, so that 
they are constantly breaking off or collapsing, only to become still more readily the prey 
of the borers, which riddle them through and through, so that they finally break down 
into small fragments and sand, The sand-feeding organisms are the last to act on the 
fragments. Holothurians, Enteropneusts and others swallow them, consuming their organic 
matter, ultimately to extrude them worn down into still smaller fragments. The carbonate 
of lime is dissolved and removed by solution, and the silica and other insoluble matters 
are left as a fine mud or clay. Currents further play their part, and indeed in North Mahlos 
no clay or mud is found, the whole of the bank being swept almost clean, 
The breaking down of the surface of any lagoon reef, together with the coral growth, 
keeps its appearance almost constant, but the reef is seldom or never still. Its surface-flat 
is everywhere either growing outwards, or, owing to beimg broken down and eroded outside, 
retreating gradually, if the reef be a faro, on its velu. If the foundations for the flat are 
being built up outside, it naturally extends, but the corals of its inner part are killed 
by the fine mud, the product of the boring organisms, swept over from outside. A few 
colonies of Porites alone survive for some time, but even these finally go, dying as soon 
as mud begins to cling to them or silt them up. From the edge of reef-flat of any reef 
in the more enclosed parts of the lagoon of North Mahlos the bottom falls gradually to 
3 or 4 fathoms, and then precipitously to some depth within 7 or 8 fathoms of the 
bottom in the vicinity. Far more of the coral from the reef above is swept outwards 
than over the reef and into its velu, if present; indeed the sandy flat, that is found 
commonly inside such reefs, is generally bare and devoid of fragments of dead corals, save 
of those few species which may very occasionally be observed growing in such _ positions, 
This being the case, it might be expected that these fragments would be swept back 
over the steep and so gradually extend the reef outwards on all sides. That this is often 
what happens is quite clear in some parts of Miladumadulu, but it depends on several 
factors, whether this is generally so in any atoll. The first in importance of these is the 
circulation of the water, which, if not free, will not remove by solution the excess of 
carbonate of lime thrown down by the organisms. Further the boring and _ sand-eating 
forms, which play a most important part, are not nearly so numerous in open as in 
enclosed lagoons. An example of such an enclosed bank is Addu, where in the lagoon 
the reefs are growing up, and sand, formed from them, is gradually raising the level of 
the bottom. As a contrast Haddumati and Kolumadulu have an ample supply of water, 
and the interior reefs save near passages have absolutely precipitous sides (on which the 
lead finds no resting-place) from 2 or 3 fathoms to near the bottom. At the foot of the 
steep are found masses of dead and decaying coral with a very occasional living colony, 
but on the steep itself there is no growth of coral nor of other organisms. In an open 
bank such as Miladumadulu there is no precipice, but merely a steep slope commencing 
