33 
a Study in Taxonomic Distribution . 
the species predicted for Stewart that are not recorded. Under prediction 
A everything was exact. Under B (species reaching the Chathams, Lake 
Taupo, and Fovea ux Strait) the following have not been found in Stewart, 
though predicted as likely : 
Plagianthus divaricatus (cf. p. 42) 
Geranium dissectum 
Sophora tetraptera 
Potentilla anserina (cf. p. 42) 
Epilobium pallidiflorum (cf. p. 42) 
„ chionanthum 
„ insulare (cf. p. 42) 
Mesembryanthemum australe (cf. p. 42) 
Hydrocotyle moschata 
Oreomyrrhis andicola (cf. p. 42) 
Daucus brachiatus 
Coprosma robusta 
„ Cunninghamii 
Erechtites quadridentata (cf. p. 42) 
Lobelia anceps 
Wahlenbergia gracilis (cf. p. 42) 
Myosotis spathulata (cf. p. 42) 
Calystegia tugoriorum (cf. p. 42) 
Dichondra repens 
Solanum nigrum 
„ aviculare 
Myoporum laetum 
Mentha Cunninghamii (cf. p. 42) 
Salicornia australis 
Polygonum serrulatum 
Pimelea arenaria 
Scirpus frondosus (cf. p. 42) 
„ americanus 
Deyeuxia Billardieri (cf. p. 42) 
These are twenty-nine in all, of which sixteen are wides, and there is no 
reason to suppose that their absence from Stewart is due to anything other 
than the fact that Foveaux Strait was formed before they reached so far 
south in their wanderings. It is noteworthy that the list contains only 
three Monocotyledons, though this group forms so large a proportion of the 
flora of Stewart (111 of 31 1). 1 
Finally, under prediction C (Auckland species reaching to Foveaux 
Strait and the north end of South Island) there were wrongly predicted : 
Colobanthus Billardieri (cf. p. 42) Rostkovia gracilis 
Epilobium confertifolium 
to which the same explanation may apply. 
General Features and Relationships of the Flora. 
We shall now go on to apply the method of prediction in greater detail, 
with the view of bringing out the general features and relationships of the 
flora of Stewart. 
(1) As the family is older than the genus, the genus than the species, 
one will expect that in comparison with the flora of New Zealand proper 
a greater proportion of families will be represented than of genera, of 
genera than of species. Verification of this gives : 
1 In a later paper I hope to deal with the distribution of the Monocotyledons, which shows 
many interesting points, e. g. a regularly diminishing percentage from one end of New Zealand to the 
other, and twice as great a percentage in the Aucklands as in the Kermadecs, with an intermediate 
figure in the Chathams, and so on. 
D 
