210 rniversiti/ of California Fuhlications in Botany [Vol. 6 



to right, or from above downward, leavas no other inference possible. 

 It is true that there are some important exceptions to this regularity ; 

 and these, as will be shown, possess a positive meaning; but they 

 acquire this significance only with reference to the general trend of 

 relationship as based on absolute number of species. It would be 

 po.ssible to establish an exact, quantitatively expressed correlation 

 between tlie numbers of species on the various islands and the num- 

 bers of species possessed jointly by them with any given island. The 

 range of floral wealth, however, is great, and the series are small ; so 

 that the degree of their trend, and its uniformity, are readily visible 

 without more elaborate mathematical treatment. The point of the 

 basic importance of riclmess of flora can therefore be accepted as 

 established without further analysis or discussion ; and I will proceed 

 to examine briefly for each island the meaning of the departures from 

 normal tendency of its series. 



ALBEMAELE 



The fairly considerable though secondary effect of geographical 

 nearness is evident from the first three figures in Table VI. Nearly as 

 many Albemarle species have been found on nearby James as on more 

 remote Charles and Chatham, though these are almost half as rich 

 again in total species as James. Narborough and Hood point the 

 same moral. Their total species number the same — eighty and seventy- 

 nine, to be exact ; those which they share with Albemarle are, however, 

 sixty-nine and fifty-eight. Hood, however, is about a hmidred miles 

 distant, while Narborough is separated from Albemarle by only a 

 narrow channel, and moreover is shielded by it from all the remainder 

 of the group. In view of this location it might be presumed that the 

 difference between the Narborough and the Hood identities would 

 be much greater: evidently position, while a factor, is not the pri- 

 marily determining one. 



Five islands have a flora of nearly the same size : Seymour with 



fifty-two species, Barrington with forty-eight, Gardner near Hood*^ 



with forty-eight, Bindloe with forty-seven, Jervis with forty-two. Their 



forms held in common with Albemarle are respectively thirty-four, 



twenty-nine, thirty -five, thirty-six, and thirty-two. The one siguifi- 



Dr. Stewart gives figures also for Gardner near Charles, but as the number 

 of species reported from this island is minimal, I have omitted all reference to 

 it. Professor Robinson mentions only "Gardner Island," and treats it as if 

 near Charles, but the number of species attributed by him to it shows that his 

 data probably pertain to Gardner near Hood. 



