NO. 3 dunkle: plant ecology, channel islands 315 



reported from Guadalupe. All of the thirty-five genera common to the 

 two islands are to be found in California. It thus seems logical to con- 

 clude that the flora of Guadalupe Island shows a greater affinity with 

 the Channel Islands than with Cedros Island. 



The Guadalupe flora appears to be composed of three elements: (1) 

 a large Mexican component, of southern California affinity, part of 

 which is also on Cedros; (2) a large number of ancient Catalinia relicts 

 not on Cedros; and (3) a California element, with many plants of 

 northern affinity, common to the Channel Islands and mainland Cali- 

 fornia but not present on Cedros. Insular endemics and Catalinia relicts, 

 common particularly to Guadalupe and San Clemente and their com- 

 mon volcanic nature, indicate the possibility of former land connection. 

 Migration of plants by the agency of birds would be possible but it is 

 difficult to understand why they would so exclusively favor the Guada- 

 lupe-Channel Islands route to that of the Guadalupe-Cedros. There 

 is a southeastern drift of ocean water along the edge of the continental 

 shelf that might possibly carry drift from the Channel Islands to Guada- 

 lupe but with the tendency of the shelf waters to drift shoreward there 

 is at least an equal chance that plants would be carried to Cedros. In 

 this connection it is not probable that seeds or plants of any of the land 

 species would be able to survive ocean transportation for the distance 

 involved. 



In view of the trend of the edge of the continental shelf, the exten- 

 sive banks and submarine ridges lying south and southwest of the Chan- 

 nel Islands, and the great epeirogenic activity which has taken place in 

 this area during past geologic periods, it seems simpler to assume that 

 there may have been land connection sometime between the two re- 

 gions. Definite geological evidence is, at present, insufficient to lend 

 more than nominal support to this assumption. However, in view of 

 the trend of the submarine ridges, the apparent absence of transverse 

 ranges in Baja California, and the trend of the ''Santa Ana Embayment" 

 of the Miocene (Reed, 1933), the suggestion is made that if a land 

 connection ever existed it would most probably have been in the nature 

 of a peninsula reaching south from Catalinia, possibly in the Vaqueros 

 formation of the Miocene. Such a suggestion has been incorporated in 

 the accompanying map of Catalinia, figure 12. 



Since the time this study was made there has been considerable 

 oceanographic, taxonomic, and geologic work on the areas including 

 Cedros, Guadalupe and San Clemente. While the author has not been 

 able to review this thoroughly the general impression gained has tended 

 to substantiate these conclusions. 



