37 
a Study in Taxonomic Distribution. 
Thus all the endemics are to be found in the to largest families in New 
Zealand, and as a matter of fact in 6 of them, which contain in New Zea- 
land 544 species (total for New Zealand, 91 families and 1,301 species), and 
in Stewart 137 species (total for Stewart, 54 families and 311 species). They 
belong, that is, to the most ‘successful’ families in New Zealand. The 
great proportion belonging to Compositae is noticeable, and goes to indicate, 
with other parallel evidence that could be produced, that species are some- 
what readily formed in this family. 
(8) In the same way, one will expect the Stewart endemics to belong 
on the whole to the larger genera of New Zealand, and testing this, we find 
that Olearia has 35 species, Raoidia 17, Cotida 19, Abrotanella 7 (the 
nearest approach to the average of 4*2), Senecio 30, Carex 54, Danthonia 13, 
Ligusticum 18, Ranunctdus 38, and Dracophyllum 18, a total of 249 species 
for 10 genera, or practically 25 each, almost six times the average for 
a genus in New Zealand. 
Relationships of the Flora to those of the Outlying Islands. 
We shall now go on to deal with the relationship of the flora of 
Stewart to those of the Kermadecs, Chathams, and Aucklands, which (see 
maps) are islands outlying much farther from New Zealand proper, but at 
the same time islands that must have received the bulk of their flora either 
by way of New Zealand, as in the case of the Chathams more especially, or 
at any rate from the same invasions, as in the case of the Kermadecs (to 
some extent) and the Aucklands. As authority for these floras I shall 
continue to use Cheeseman’s Flora ( 1 ), supplemented by his later lists in 
Chilton’s ‘ Subantarctic Islands of New Zealand ’ ( 2 ). 
Stewart Island lies in lat. 47 0 S., the Aucklands in 5 1°, the Chathams 
in 44 0 , and the Kermadecs in 30°. The distance in a straight line from 
Stewart to the Aucklands is about 200 miles, to the Chathams about 700 
miles, and to the Kermadecs about 1,200 miles, including nearly the whole 
of New Zealand, which lies between. Stewart is chiefly archaean rocks, the 
Aucklands all of igneous origin (some parts granites and gabbros, but 
mostly volcanic), the Chathams schists, volcanic rocks, or tertiary sediments, 
and the Kermadecs entirely volcanic ( 4 ). 
(9) It is clear from the maps, especially the second map on p. 26, that 
while the plants of Stewart Island are old in New Zealand, those of the 
other more outlying islands will be older, in general, whether they were 
derived from New Zealand, or whether they reached these islands on the 
way to New Zealand. One will therefore expect to find that Stewart has 
a large proportion of species in common with these different islands, and 
that they have a larger proportion of species in common with Stewart than 
they have with New Zealand as a whole, though in general they are nearer 
to New Zealand. Testing this, we find : 
