5 1 6 Guppy. — Testimony of the Ende 7 nic Species of the Canary 
This brings us to another stage in the argument. We now perceive 
that whilst the species peculiar to and characteristic of the Macaronesian 
floral region have a much wider range in the Canaries than the purely 
Canarian endemics, they are also far older. Whilst the Macaronesian 
plants are remnants of old Tertiary forests of Southern Europe and North 
Africa, and represent types that have disappeared from the continent, the 
purely Canarian endemics belong to types still predominant in the Mediter- 
ranean region. The Macaronesian endemics are true ‘relicts’ and are 
widely spread over the Canarian archipelago ; whilst the purely Canarian 
species are of recent and local origin, and are for the most part limited to 
single islands. As already implied, the species peculiar to the Canaries are 
eight times as numerous as the Macaronesian species. The purely Canarian 
species would, in accordance with the theory of Dr. Willis, be the most 
liable to extinction, and this finds support in the behaviour of the Statices 
of the subsection Nobiles as described by Dr. Stapf (‘ Annals of Botany xx, 
xxii). They are very local and ‘ there is a considerable risk of their total 
disappearance ’. We should not look for the same with the Macaronesian 
species. But they must go the way of all plants, and one of them ( Clethra 
arbor eci) has not been found since 1828. 
In the more recent Flora of the islands by Pitard and Proust ( 1908 , 
p. 77) we have important light thrown on the origin of new species in the 
Canaries. If they had not been following the practice long in vogue among 
systematists in dealing with insular floras, a practice well illustrated in the 
quotation from Hooker (already given) when he was discussing the origin 
of the Canarian endemic species thirty years before, one might have credited 
the authors with anticipating Dr. Willis in the matter of the Age and Area 
theory. But, as held by the writer, the real significance of this theory lies 
in its return to a pre-Darwinian position respecting plant distribution. 
Dr. Willis has here rendered the greatest service to botanical geography by 
demonstrating the importance of principles that had been almost forgotten 
in the efforts to apply the great theory of Darwin to the central problems 
of the plant world. 
Viewed from the standpoint of Age there were for Pitard and Proust 
two types of endemic species in the Canaries. There were in the first place 
those found generally distributed over the group, very ancient forms that 
once existed in the neighbouring continent, but now survive only in the 
Macaronesian region. (These are the Macaronesian endemics before 
recognized.) Then there were the much more numerous recent forms, mostly 
localized in single islands and derived, just as Hooker held, from parent 
forms already in the archipelago. These are the purely Canarian species, 
and the authors of this Flora make some very suggestive remarks on 
Nature’s mechanism in their production. 
They take the case of the representatives of Micromeria , a cosmopolitan 
