New Zealand Flora , with a Reply to Criticism. 363 
some other hypothesis, which must be mechanical, owing to the fact that 
the figures show such mechanical regularity. 
Dr. Sinnott bases his views largely on the undoubted fact that the 
proportion of ‘ swamped 5 genera is larger in the more outlying of the big 
islands—in New Zealand than in Ceylon, in the Hawaiian islands than in 
New Zealand. But that mere isolation is not sufficient as an explanation 
would seem to show in the fact that in the very isolated islands round New 
Zealand the proportion is not so high as in New Zealand itself. In New 
Zealand 151 genera out of 316 show it, in the Kermadecs only 8 out of 62, 
in the Chathams the same, and in the Aucklands 12 out of 64. In none of 
the islands is the proportion anything like so high as in New Zealand, and 
it is highest in the Aucklands, which were probably nearest to the incoming 
stream of plants. On the other hand, the number of genera which are 
swamped in Neiv Zealand is 13 in the Kermadecs, 33 in the Chathams (the 
most isolated), and 26 in the Aucklands, facts tending to show that the 
swamped genera were in existence fairly early ^opposite to the Chathams, 
and therefore were rather old in comparison to some of the rest, though even 
in the Chathams the unswamped genera are almost as numerous (29). 
Another test that we may apply is to find the proportion of ‘ swamped ’ 
genera in the northern and southern invasions of plants above discussed. 
The northern shows 45 out of 75 or 60 per cent., while the southern shows 
36 out of 108 or 33 per cent. We have seen that probability is in favour of 
the greater age in New Zealand of the northern invasion, so that to some 
extent this speaks in favour of Dr. Sinnott’s views, in a general and purely 
local sense. But as only one herb ( Elatostema ) is swamped in the northern 
invasion, and all the shrubs but one ( Veronica ) in the southern, it is, it 
seems to me, equally possible that swamping may go with woody habit, and 
further tests are necessary. 
Of the 151 ‘ swamped ’ genera, 45 Dicotyledons and 30 Monocotyledons 
are herbs, or 50 per cent., while of the 165 unswamped, 75 Dicotyledons and 
63 Monocotyledons are herbs, or 83 per cent. Of these unswamped genera 
99 have no endemics, and of these 85,01* 85 per cent., are herbs, while of the 
66 with endemics 53, or 80 per cent., are herbs. From these figures it would 
seem that the evidence is just as good for the connexion of swamping and 
woody nature as of swamping and age. 
The Coniferae are probably older than the flowering plants, and as 
they have no wides at all in New Zealand, this speaks in favour of age, but 
they are also all woody plants. The Ferns, on the other hand, which are 
probably older again, show very little 4 swamping’, only 5 genera out of 
3 i exhibiting this phenomenon. Of these it may be noted that three are 
the only tree-ferns in New Zealand. The remaining two, and all the un¬ 
swamped genera, are herbaceous. It is evident that the question of swamp¬ 
ing must be first disentangled from the question of the relatively greater 
