have done so, but I think not. There is, I think, every probability that
most islands, etc., where a varied fauna now exists have been once more
extensive, e.g. New Zealand, Madagascar. Where there is no such evidence
(e.g. Galapagos), the fauna is _very restricted_.

Lastly as to want of references; I confess the justice of your
criticism. But I am dreadfully unsystematic. It is my first large work
involving much of the labour of others. I began with the intention of
writing a comparatively short sketch, enlarged it, and added to it, bit
by bit; remodelled the tables, the headings, and almost everything else,
more than once, and got my materials into such confusion that it is a
wonder it has not turned out far more crooked and confused than it is.
I, no doubt, ought to have given references; but in many cases I found
the information so small and scattered, and so much had to be combined
and condensed from conflicting authorities, that I hardly knew how to
refer to them or where to leave off. Had I referred to all authors
consulted for every fact, I should have greatly increased the bulk of
the book, while a large portion of the references would be valueless in
a few years owing to later and better authorities. My experience of
referring to references has generally been most unsatisfactory. One
finds, nine times out of ten, the fact is stated, and nothing more; or
a reference to some third work not at hand!

I wish I could get into the habit of giving chapter and verse for every
fact and extract, but I am too lazy and generally in a hurry, having to
consult books against time when in London for a day.

However, I will try and do something to mend this matter should I have
to prepare another edition.

I return you Forel's letter. It does not advance the question much,
neither do I think it likely that even the complete observation he
thinks necessary would be of much use; because it may well be that the
ova or larvae or imagos of the beetles are not carried systematically by
the ants, but only occasionally owing to some exceptional circumstances.
This might produce a great effect in distribution, yet be so rare as
never to come under observation.

Several of your remarks in previous letters I shall carefully consider.
I know that, compared with the extent of the subject, my book is in many
parts crude and ill-considered; but I thought, and still think, it
better to make _some generalisations_ wherever possible, as I