70 AGE AND AREA [pt. i 



and only 19 in the next class, which does not in all cases include 

 Stewart. In other words, most of these species were so old that 

 they were also in time to reach Stewart before it was cut off. 

 The endemics that reach the islands also show 41 in class 1 and 

 21 in class 2, but the wides that do not reach these islands 

 (last column in table above) show 35 in class 1 and 39 in class 2, 

 indicating that they were on the whole a good deal younger, so 

 that many of them were not in time to reach Stewart. The 

 endemics that do not reach the islands show 52 and 60 in these 

 classes respectively, in the same way. 



That these outlying islands of New Zealand are not a special 

 case may be seen by comparing with the flora of Great Britain 

 those of some of its outlying islands. If we take the Orkneys 

 (north Scotland), Colonsay (south-west Scotland), Clare (west 

 Ireland) and the Scillies (south-west England), islands widely 

 separated, and differing very much in climate and geology, 

 and if we take in these, at random (37, 108), the families 

 Ranunculaceae,Caryophyllaceae, Leguminosae, Orchidaceae, and 

 Gramineae, we find that Avhile (going by the LondoJi Catalogue^ 

 8th ed.) the average distribution of a species in Great Britain 

 is to 47 of the vice-counties out of 112, the 175 species of these 

 families that occur on the islands mentioned range on an average 

 to 71 (or 50 per cent, more), whilst those that reach three or four 

 of the islands show an average range of 99. The facts are exactly 

 parallel to those for the islands off New Zealand, though of course 

 not so striking, as the islands are very much closer to their 

 mainland. 



Before going further we must once more consider the reserva- 

 tions which are laid down in the statement of the hypothesis 

 in the preceding chapter, and whose misunderstanding seems the 

 cliief stumbling-block in the way of an acceptance of Age and 

 Area. It is easy to pick out of the list of "wides" reaching the 

 islands a few that have less range in New Zealand than other 

 wides that occur there and do not reach the islands. The hypo- 

 thesis is often treated in this manner, and then rejected for non- 

 agreement with actuality. It must not be forgotten that if it 

 could be applied in such minute detail we should have at our 

 command a theory that would (explain more facts in distribu- 

 tion and phylogeny than any other that has ever been suggested. 

 Too much is expected of an hypothesis which claims no more 

 than to be a useful guide, and the reservation, that it must not 

 be applied to a group of less than ten allied species, is ignored. 



