448 Willis . — The Distribution of Species in New Zealand . 
explaining it away, or of saying that the progression of the numbers is 
accidental. In only 12 families out of 70 is there any break in the regular 
progression to a single maximum, and then only in 16 figures out of 770. 
In any case the greatest ‘ error ’ is only 4, but in the prediction we have 
shown that two maxima may at times occur. 
A very brief consideration of these tables is enough to show that 
natural selection is an agent quite incapable of causing such a simple 
arithmetical arrangement. The area occupied by a species cannot be due 
to it, though it no doubt enters to a large extent in determining the 
commonness of a species within its area of occupation. 
The complete and surprising manner in which these tables bear out the 
prediction made (and be it noted that I made it first and verified it after- 
wards, thus making a discovery about the grouping of plants in New 
Zealand which had hitherto remained unmade) gives great support to the 
hypothesis of ‘ age and area 5 which I based upon the figures of the 
Ceylon flora. 
A study of the diagram just given leads to another prediction which 
may be made. It will be seen that on the average the species with the 
longest range in the country have the lower points of their triangles towards 
the outer ends of the islands. Thus the only species that have one 
boundary in the first occupied zone, from 100 to 300 miles, have ranges in 
the country of 800, 600, and 400 miles (average 600), and those that have 
a boundary in the last zone, 900-1,000 miles, have ranges of 800 and 
600 miles (average 700) ; whereas those with boundaries in 400-500 miles 
have ranges of 200 and 200, and those in 500-600 have ranges 400 and 200 
(average 300). As with the preceding prediction, cases may occur in which 
there may be two maxima, but on an average, and on the total, one will 
expect to find, assuming ‘age and area’ to hold, that the ranges of the 
species will be greatest in the case of those commencing in the outermost 
zones, and least in the case of those commencing towards the centre. This, 
it need hardly be remarked, is quite inconsistent with natural selection. 
Turning now to the actual facts, one finds (taking the grand total of 
the flora) that the range of those endemic species which have their 
boundaries in the different zones is as follows : 
Table VII. 
Boundary (N. or S., or both). 
Average possible range} 
Actual range. 
Between 0 and 200 miles 
980 or 100 miles. 
691 miles. 
„ 201 ,, 400 ,, 
780 „ 300 „ 
5^8 ,, 
„ 401 „ 600 „ 
5 ) 5 °° )i 
321 „ 
„ 601 „ 800 „ 
380 „ 700 „ 
272 „ 
„ 801 „ 1,000 „ 
„ 1,001 „ 1,080 „ 
l80 ,, 900 ,, 
463 „ 
40 ,, 1,040 „ 
753 „ 
1 If the species ranged indifferently both ways, and not chiefly towards the centre, the average 
range should be the same at every point. On the theory of natural selection there is no particular 
reason why they should range one way more than another. 
