Willis . — The Distribution of Species in New Zealand. 439 
animals 1 — convinced me that such confirmation was easily to be obtained, 
and I cast about for a flora from which to procure figures as nearly like the 
Ceylon figures as was possible, deciding 2 for that of New Zealand, which for 
this purpose has many advantages. In the first place, it provides a test of 
endemism and of anything concerned therewith which by general admission 
has almost no rival in the world. Ceylon is not mentioned in Wallace’s 
‘ Island Life’, whereas almost three chapters are devoted to New Zealand. 
In the second place, the flora is well known, and it is very unlikely that 
further exploration will seriously alter the areas in which the various species 
are known to occur. 3 In the third place, the islands have no sudden or 
violent changes of climate in any part which might form a barrier to 
migration ; and lastly, they are very convenient for determining the area 
occupied by any species, for they are spread out in a long curve running in 
general north and south for approximately 1,080 miles, 4 with an average 
breadth of about 100, 5 so that the area occupied by a species can be, and for 
the purposes of this work has been, roughly estimated by the distance between 
its extreme northern and southern limits ; or rather, it would be more correct 
to say that longitudinal range is employed here in place of actual area. The 
map of New Zealand was marked by transverse lines at every twenty miles, 
which rendered determination simple. This method does not give close 
accuracy, but the islands are so narrow that the error is small, and on the 
total it cancels out. In a few instances the result was to place in the last 
class (range, 1-40 miles) species which ranged crosswise to the islands to such 
a distance that they should have gone into the next class above (41-160 m.). 
As a matter of fact, as will be seen, the figures are so conspicuously in 
accord with my hypothesis of ‘ age and area ’ that they would not be 
vitiated by a source of error considerably larger. 
For the purposes of this work I have taken the flora as being that 
given by Cheeseman in his ‘ Manual of the New Zealand Flora ’, pub- 
lished at Wellington in 1906, and I have absolutely neglected all later 
work. In this way the facts upon which I found my conclusions are 
available to any one without difficulty, and, as I have explained above, 
detailed confirmation work does not affect the general results. 
Cheeseman, however, includes in his Flora not only the plants of 
1 By the kindness of Prof. J. S. Gardiner, F.R.S., I have been able to go over certain groups of 
animals whose statistics of classification and distribution he considers to be reliable, and in some 
of these, which I have actually enumerated, I find my age and area hypothesis borne out. 
2 Without making any preliminary investigations whatsoever. Several botanists had said to me, 
‘ Ceylon is all very well, but is not really a test of endemism : try New Zealand 
3 The general effect of detailed completion work, such as most of that published about New 
Zealand in the last ten years (since Cheeseman’s Flora), is to add new species to the lowest class of 
all and to raise other species, especially in the two lowest classes, in the scale, but to leave the 
general result unaffected. 
4 1, 1 10, if the Three Kings islets at the north end be included. 
5 In actual fact approximately 99*5 miles. 
H h 
