124 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
nesium carbonate, and consequently to form a dolomite from a rock of 
this character would require a very extensive solution and concentration 
of the original limestone. The experiments of Klement, in 1895, dem- 
onstrated that the unstable mineral aragonite is much more susceptible 
of the invasion of magnesium carbonate than the stable form calcite. 
The conditions under which the experiments were performed make a 
strict comparison with reef dolomitization difficult. The results are, 
however, very suggestive as showing that magnesium sulphate can, under 
certain conditions, react with calcium carbonate. The stable compound 
calcium magnesium carbonate is deposited, while the by-product calcium 
sulphate can be kept in solution by dilution. It may be that the reac- 
tion which proceeds quickly under conditions of great concentration and 
high temperature may proceed slowly, but no less surely, in superficial 
waters such as those of a lagoon which is maintained for a sufficiently 
long period of time under uniform conditions. 
Recently it has been suggested that the introduction of magnesium 
into a coral limestone is only effected when the rock has been submerged 
for some time to a considerable depth corresponding to a particular pres- 
sure. This view is, however, not in harmony with the facts at Christmas 
Island, Mango, Vatu Vara, Ngillangillah, etc., where the highest rocks of 
the island are dolomitized, and the only movement of which there is 
evidence since their formation in shallow water is one of elevation. It 
is true that denudation has removed a certain thickness of rock since 
their elevation, but there is evidence that in the islands mentioned the 
effects of denudation have not been extensive. 
It will be seen that most of the opinions which have been put forward 
to explain the introduction of magnesium carbonate into coral rocks do 
not agree with the facts herein recorded as to the distribution of dolo- 
mite in upraised coral islands. There yet remains Dana’s well-known 
theory of dolomitization, and perhaps some modification of it will be 
found to harmonize more closely with the evidence than any view which 
has, as yet, been put before geologists. The occurrence of dolomite at 
the summits of many of the islands is significant. At Christmas Island 
dolomite is found near the summits of the highest hills immediately 
below beds of phosphate. Dr. C. W. Andrews believes that after 
Miocene times the movement of subsidence to which the formation of 
the older limestone is due, ceased, and a very long period of rest ensued, 
during which the only land masses rising above the lagoon were the low 
islets now forming Murray, Ross, and Phosphate Hills. The accumula- 
tion of guano from the droppings of sea birds, and its subsequent altera- 
