_Frith Hill, Godalming. August 26, 1888._

Dear Mr. Hemsley,--You are aware that Patrick Geddes proposes to exclude
Natural Selection in the origination of thorns and spines, which he
imputes to "diminishing vegetativeness" or "ebbing vitality of the
species." It has occurred to me that insular floras should afford a test
of the correctness of this view, since in the absence of mammalia the
protection of spines would be less needed.

Your study of these floras will no doubt enable you to answer a few
questions on this point. Spines and thorns are, I believe, usually
abundant in arid regions of continents, especially in South Africa,
where large herbivorous mammals abound. Now, if the long-continued
presence of these mammals is a factor in the production of spines by
Natural Selection, they should be wholly or comparatively absent in
regions equally arid where there are no mammals. The Galapagos seem to
be such a case--also perhaps some of the Sandwich Islands, and generally
the extra-tropical volcanic islands. Also Australia comparatively, and
the highlands of Madagascar.

Of course, the endemic species must be chiefly considered, as they have
had time to be modified by the conditions. If you can give me the facts,
or your general impression from your study of these floras, I shall be
much obliged. I see, of course, many other objections to Geddes's
theory, but this seems to offer a crucial test.--Believe me yours very
truly,

ALFRED R. WALLACE.