502 
Willis . — Endemic Genera of 
which even in these isolated floras are waging a losing fight against more 
vigorous and adaptable new-comers.’ 
This explanation has always been unsatisfactory to me since I began to 
study in situ the numerous endemic forms of Ceylon, and a great argument 
’ against it, which has always been’ passed over and left unanswered by its 
supporters, is the fact that endemic species usually occupy continuous areas. 
In other words, from the point of view of the current explanation, the 
endemics have not retired in confusion, but have kept their formation. 
Other arguments which also turned up in the course of my work confirmed 
me in this disbelief of the usual explanation. For example, I found it very 
difficult to reconcile with the idea that endemics were in general moribund 
the fact that in a single country there was a regular graduation of endemics 
upwards from many of small area of distribution to few of large, and of 
species of wider distribution than endemic in the opposite direction 
(cf. tables in 14 , p. 310; 15 , p. 3; 16 , p. 44^,; 19 , pp. 336, 338, 344). 
Again, practically all the endemics of a country, if they possess more than 
one or two species, show graduated maps like that given for Ranuncidus in . 
New Zealand, and the same is the case whether they have or have 
not widely distributed species of the same genus beside them. Another 
awkward point for the supporter of generally moribund endemics (i. e. other 
than a comparatively few, too small in number to affect the figures) is the 
fact that in the far outlying islands round New Zealand (Chathams, &c.) the 
more widely distributed species there are, the more endemics ( 18 , p. 332, 
Table III), and the wides are more numerous in proportion (figures in 20 , 
bottom of p. 352). Further, the endemics of New Zealand and Ceylon are 
most numerous where there are most wides, least where there are fewest, 
which is not at all what one would expect if the endemics are moribund 
species. The comparatively few endemics found near the outer ends of 
New Zealand range over a much greater area than do those in the middle 
(cf. the map of Ranunculus above, and table in 16 , p. 448). Yet another 
great difficulty, from the old point of view, is the fact that species endemic 
to New Zealand and its immediately outlying islands (Kermadecs, 
Chathams, Aucklands) are on the average more widespread in New Zealand 
than the species common to New Zealand and the outside world, but not 
found on these little islands, while the species found in the outside world, 
and occurring also in New Zealand and these islands, are the most wide- 
spread in New Zealand of all. Fern endemics, which must on the whole 
be older than angiosperm endemics, occupy more area than the latter. 
Lastly, and perhaps most important of all, the endemic species belong 
principally to the large , or what on the old view are considered the 
successful genera, and much less, even in proportion, to the small. 
All these, and many other arguments which might be brought forward 
(cf. 20, p. 352), rendered it impossible for me any longer to adhere to the 
