278 Willis . — The Floras of the Outlying Islands of 
MYOSOTIDIUM , Veronica (4), Plantago , A scar in a, Urtica , Bulbinella , 
Rostkovia , Carex (2), Imperata , Hierochloe , Deschampsia (2), Poa (7), Festuca 
(2). Those which have no number against them are represented by one 
species only, and those in capitals are endemic genera of the islands only. 
No one can look over this list of genera, and honestly suggest that it 
represents a set of relicts. Many are genera of world-wide distribution, 
and all are common, except the three endemic genera, with 5 species, which 
represent most of the species which might perhaps be looked upon as 
relicts. 
(24) We have seen (10, pp. 358, 361) that the northern invasion of New • 
Zealand was nearly all tree-like, the southern herbaceous. We shall there- 
fore expect the proportion of trees to be greatest in the endemics of the 
northern islands, of herbs greatest in the southern. In actual fact, the 
Kermadecs have 7 endemic shrubs and trees and 2 herbs (both Mono- 
cotyledons), the Chathams have 13 shrubs and trees and 12 herbs (4 Mono- 
cotyledons), and the Aucklands 3 shrubs and trees only against 27 herbs 
(7 Monocotyledons), whilst in the Campbells there are 3 shrubs, in the 
Antipodes 1, and in Macquarie none. This bears out the prophecy com- 
pletely, and it may also be pointed out that the three endemic genera of the 
islands are composed entirely of herbs ( Stilbocarpa , Pleurophyllum , Myo- 
sotidiu 7 fi) } a fact which does not harmonize well with Prof. Sinnott’s suggestion 
that endemics are relicts and herbs the youngest forms. 
It is clear, from the success of the 24 predictions above given, that, just 
as in the case of Stewart Island, it is hardly conceivable that the out- 
lying islands should have received their floras by casual transport across 
the intervening seas. There must have been connexion by land at some 
time. When one works out their species one by one, it is fairly clear that 
a few — at most hardly exceeding, if they even equal, 10 per cent. — may 
have been carried by water or by birds, but the great bulk must undoubtedly 
have arrived by land. In no other way could the great similarities between 
the islands, or between the islands and Stewart, have come about. No 
casual transport could give results like those above, nor would it occur that 
the island endemics belonged almost solely to the larger families. 
The Kermadec Islands. 
We must now go on to consider in brief some of the relations of the 
islands to one another, and to New Zealand and its probable invasions of 
plants (10, p. 355). Let us begin with the Kermadecs. They have 
altogether, according to Cheeseman, a flora of 71 species, made up thus: 
Table XIV. 
Dicotyledons 31 families 47 genera 56 species 
Monocotyledons 6 „ 15 ,, 15 ,, 
