possibility of New Zealand having once been united with Australia (p.
446, 4th Ed.). The puzzle is greater than I can get over, even looking
upon it as an oceanic island. Why should there have been no mammalia,
rodents and marsupials, or only one mouse? Even if the Glacial period
was such that it was enveloped in a Greenlandic winding-sheet, there
would have been some Antarctic animals? It cannot be modern, seeing the
height of those alps. It may have been a set of separate smaller
islands, an archipelago since united into fewer. No savages could have
extirpated mammalia, besides we should have found them fossil in the
same places with all those species of extinct Dinornis which have come
to light. Perhaps you will say that the absence of mammalia in New
Caledonia is a corresponding fact.

This reminds me of another difficulty. On the hypothesis of the coral
islands being the last remnants of a submerged continent, ought they not
to have in them a crowd of peculiar and endemic types, each rivalling
St. Helena, instead of which I believe they are very poor [in] peculiar
genera. Have they all got submerged for a short time during the ups and
downs to which they have been subjected, Tahiti and some others having
been built up by volcanic action in the Pliocene period? Madeira and the
Canaries were islands in the Upper Miocene ocean, and may therefore well
have peculiar endemic types of very old date, and destroyed elsewhere. I
have just got in Wollaston's "Coleoptera Atlantidum," and shall be glad
to lend it you when I have read the Introduction. He goes in for
continental extension, which only costs him two catastrophes by which
the union and disunion with the nearest mainland may readily be
accomplished.... --Believe me ever most truly yours,

CHA. LYELL.