1916] Kroeher: Floral Relations Among tlie Galapagos Islands 205 



The situation is still clearer if we regard only the nine larger 

 islands. In Table IV these are given, arranged by geographical 

 groups, each followed by the three islands with which its mean per- 

 centage relation is highest according to Table III ; and each of these 

 three names is followed by a number from 1 to 8, indicating the 

 respective degrees of geographical proximity of these islands to the 

 one in question. Those of the same group are in heavy type. 



24 



34 



31 



Again the western group is a unit, not only in its internal relation, 

 but in the fact that the same non-western islands come next in each 

 case. Much the same holds of the northern group. The three southern 

 islands are again clearly linked together. The addition of the num- 

 bers indicating proximity points in the same direction : 24, 34, and 

 31 total 89. Divide by 27 (3 numbers each for 9 islands), the pro- 

 duct is 3.3. If, however, geographical relations did not exist, the 

 numbers would have appeared at random, and their average would 

 have been the mean of the sequence 1 to 8, or 4.5. The difference 

 between 4.5 and 3.3 is some index of the effect of geographical prox- 

 imity in increasing floral relationship between islands. 



Another and more nearly correct method of demonstrating such 

 influence of geography upon flora as there might be, subsequently 

 occurred to me. This was to arrange the islands, not in alphabetical 

 sequence as Professor Robinson had done, nor in geographical order, 

 as in Tables II, III, and IV, but in the order of their wealth of 



