74 AGE AND AREA [pt. i 



Or, lastly, one may take the endemics of New Zealand and 

 the outlying islands, and make predictions about them. We 

 have just seen that on the whole, each in its own circle, the 

 larger families and genera of a country will be the older in that 

 country. Now endemic species, by hypothesis, occupying small 

 areas, will be on the whole younger than the wides, as already 

 pointed out, and one will therefore expect the older families, 

 which have had the longest time in the coiuitry, to produce the 

 most endemics. That is to say, that the endemics should belong 

 to the largest families in the country, working in averages. The 

 same rule should of course apply to the genera. If now we test 

 this on New Zealand and its surroimding islands, we find that 

 in New Zealand and its outlying islands there are 22 families 

 above the average size, with 1100 species, of which 890 are 

 endemic to New Zealand or the islands, or 80 per cent.; there 

 are 69 families below the a^^erage, with 292 species, of which 

 only 110 are endemic, or 37 per cent., an enormous difference. 

 In Stewart Island, all the 19 local endemics belong to the 15 

 largest families of New Zealand, and ]0 of them to the three 

 largest families in Stewart, and the same thing holds for the 

 local endemics of the other outlying islands. 



In the same way, one finds that the {local) endemics of the 

 Kermadecs, Lord Howe Island, and Norfolk Island, all islands 

 which must have lain more or less in the track of the invasions 

 of New Zealand by plants from the north, belong chiefly to 

 those families and genera of their floras which have also reached 

 New Zealand, i.e. to the oldest families and genera contained 

 in them. 



On the supposition, which follows from Age and Area, that 

 the wides have given rise to the endemics (p. 61), one will expect 

 most endemics to occur in those regions where there are most 

 wides, and not, as on the theory of dying out of endemics would 

 rather be the case, in those regions where there are fewest 

 wides. In fact, this is at once seen to be the case, whether in 

 New Zealand, its outlying islands, or in Ceylon or elsewhere. 



Age and Area is thus seen to be a hypothesis by whose use 

 one may discover great numbers of new facts, and as so far all 

 the predictions made by its aid have proved to be correct, on 

 verification, the result is to afford great support to the hypo- 

 thesis itself. Over 90 such predictions as those mentioned above 

 have now been made and verified, and one may, one is inclined 

 to think, regard the hypothesis, in the absence of any rival' 



