New Zealand and their Distribution . 
289 
A cian thus S inclairii. 
Rhopalostylis sapida. 
Lepyrodia Traversii\ doubtful determination, see Cheeseman. 
Deschampsia caespitosa : see above, p. 284. 
Thus several are doubtful determinations, or coast forms, which may 
have been brought by the currents. But it is only among these few species 
that one can look for exceptions to age and area ; all the remainder of the 
flora of the Chathams — 155 species in all — is easily explained by the 
simple operation of this law, and no other assistance is needed to explain it. 
(32) We may predict that the Chathams should have proportionately 
more species in the higher classes of width of distribution than the 
Kermadecs or Aucklands, where more recent arrivals may occur, as these 
islands have been nearer to the tracks of the invasions. 
Table XXI. 
The Chathams have 155 species, of which 77 are Class 1, 30 Class 2 
,, Kermadecs „ 71 ,, ,, 22 ,, ,, 16 ,, 
„ Aucklands „ 119 „ „ 35 „ ,, 6 
The proportion (and the totals) in the Chathams is by far the greatest. 
One might give other predictions about the floras of these islands, but 
these will suffice to show that in a well-defined area like New Zealand and 
its surrounding islands one can without hesitation draw upon the hypothesis 
of age and area to make predictions about the taxonomic distribution of 
the flora, and find that the predictions are justified by the facts. I have 
now used the hypothesis to make no fewer than 67 predictions (32 in this 
paper), every one of which has proved to be correct. In several cases the 
exactness with which the prediction has been borne out by the facts has 
been positively astonishing, and in all cases the result has been as near to 
accuracy as can reasonably be expected in a biological subject, and espe- 
cially in one so complex as geographical distribution, where changes in the 
configuration of the land and sea may be continually going on, and in a very 
complicated way. It seems to me, therefore, that the hypothesis of age and 
area, having been successfully used to make so many predictions, may now 
perhaps be regarded as being fairly well established as one of the chief (if 
not the chief) positive factors in geographical distribution, while the action 
of barriers, which may be of many kinds — seas, rivers, mountains, ecological 
barriers, changes of climate, &c . — may be regarded as the chief negative 
factor, and other things, such as the action of man, are also of very great 
importance indeed. 
If this position be regarded as reasonable, it is evident that we must 
now state the hypothesis in rather more clear and definite terms than 
those in which it vyas first formulated, e. g. in 6, p, 438. After careful 
