192 NAT URALTSCIENCE. Marcu, 
‘a very extensive island,” though he doubts whether this island ever 
formed part of a continent. 
Turning to their zoology, we find that they are entirely destitute 
of mammals, but that they possess Amphibia, having two species of 
frogs and three species of Cecilia, snake-like creatures, which 
burrow underground in the manner of worms. Now, it seems impos- 
sible to explain the presence of these Amphibia unless, at some 
remote period, the islands formed part of a continent, for salt-water 
is fatal to them, and destroys even the eggs of frogs. On the other 
hand, if the former connection with a continent be admitted, what 
can be said of the absence of mammals, for one would have expected 
that some of the smaller genera, such as rats, mice, civets, lemurs, 
and insectivores would havesurvived. Dr. Wallace suggests that the 
islands have at some time been so nearly submerged that the portions 
remaining above water were too small to support the existence of the 
smallest mammals. If this be the explanation, and if we accept the 
evidence of the Amphibia as to ancient continental connection, then 
the absence of mammals in such islands cannot be taken as proof 
that they have never been part of continental land. 
New Caledonia, again, in the Pacific, is regarded by Dr. Wallace 
as anoceanic island. There are no indigenous Mammalia or Am- 
phibia, the solitary frog being known to have been introduced. There 
are several peculiar lizards and a snake (one of the Boas); and 
the island is separated from neighbouring groups by water of more 
than 1,000fathomsdeep. Notwithstanding this limited vertebrate fauna 
there is evidence that New Caledonia has once been part of an exten- 
sive land area. Stratified rocks, believed to be of Secondary and 
Tertiary age,enter into its geological structure, and there is a genus of 
land snails (Placostylus) which occurs in the neighbouring archipelagoes, 
as well as on Lord Howe’s Island and in New Zealand. Hence it 
has been recently argued that the unity and limitation of the Placostylus 
area can only be explained by the supposition that these islands are 
portions of a broken-up and submerged continent which was 
distinct from Australia. To reconcile such a supposition with the 
absence of mammals, it is only necessary to assume that this continent 
dates back to a time when mammals were not in existence in the 
Pacific region, just as Australia dates from a time when only marsu- 
pial mammals were in existence. 
The cases above mentioned show that there is no constant relation 
between the geological structure of oceanic islands and the manner in 
which they have received their present inhabitants, for there are 
islands which are not far froma continent, and have clearly formed part 
of the continental area within the limits of Tertiary time, and yet are 
without any indigenous Mammalia or Amphibia. There are other 
islands which are destitute of Mammalia and are geographically 
oceanic, but nevertheless, have a geological structure of continental 
type. In other words, there are islands which testify to the former 
