1916] Kroeher: Floral Kelations Among the Galapagos Islands 219 



two entrias in its row, as compared with western Narborough, Avhich 

 possesses an equal number of species but has no entry in its row. 



Examination of the individual figures in Table VII also shows a 

 distinct but definitely limited influence of proximity paralleling- that 

 which has been established on the basis of Table VI. Only it must be 

 remembered that in the present tabulation the high numbers miLst be 

 expected to come not near the edges of the table but along the diagonal 

 axis, dwindling away from this in all directions. A high percentage, 

 such as that of thirty-three between Hood and Duncan, is therefore 

 not indicative of operation of proximity, but is an effect of the similar 

 number of species, seventy-nine and 103, occurring on the two islands. 

 Viewing the figures with this point in mind, it could be pointed out, 

 for instance, that the percentage of species common to Albemarle and 

 James, and Albemarle and Indefatigable, is greater than the percent- 

 age common to Albemarle and Charles, and Albemarle and Chatham — 

 unquestionably as a result of proximity, since theoretical probability 

 would reverse the figures. In the same way scarcely a row" or a column 

 can be followed through without analogous deviations due to the same 

 cause. The relationships established in this way, however, follow so 

 closely those already discovered in Table VI and discussed at length, 

 that it would be mere repetition to cite and analyze them. 



CONCLUSIONS 



It follows, therefore, that, so far as the number of joint species 

 is concerned, the floras of the various Galapagos islands do not show 

 any unaccountable relations or mysteries, but almost exactly such con- 

 nections as might be expected. 



The first and fundamental element that determines the number of 

 species which two islands have in common, is clearly the number of 

 species found on each. So obvious is this both from my own tables 

 and those of Professor Robinson and Dr. Stewart, that practically all 

 of the foregoing discussion of the characteristics of individual islands 

 has concerned departures from this rule. A given island will always 

 share more of its species with an island containing 300 species than 

 with one containing fifty. This is clearly the result of the working of 

 mathematical probability, and just in proportion as the influence of 

 this element transcends that of any other, are these more specific causes 

 relegated to a subsidiary station. 



