612 ALLAN HANCOCK PACIFIC EXPEDITIONS VOL. 5 



more distant trans-Pacific islands. As in the case of the latter, the 

 Equatorial Countercurrent would appear to provide the necessary mo- 

 tility. 



In the light of this reasoning, the occurrence of the Galapagos endemic 

 species. Platypodia gemmata, at Taboga Island, Panama, in 1925 as re- 

 ported by Finnegan (1931) may be attributed to the same cause as the 

 record of the Gulf of California species, Leptodius occidentalism at Hood 

 Island, Galapagos, in the same year as reported by Boone (1927) : the 

 invasion of the Southern Hemisphere by the Equatorial Countercurrent 

 in years of exceptional disturbance such as 1891 and 1925 (cf. Murphy, 

 op. cit.). Such years would appear propitious for the establishment of 

 Galapagos species on Central and northern South American shores. 

 Their survival beyond the first generation in the new environment would 

 be the exceptional occurrence; witness the failure of Hancock Expeditions 

 to duplicate either record in the relatively stable period of 1931-39. 



While species common to the Galapagos Islands and the tropical 

 American mainland should possess Atlantic analogues unless belonging 

 to one of the classes previously enumerated, there should exist no com- 

 parable relationship between the Atlantic species and those common to 

 the Galapagos and either of the temperate faunas alone. Thus the 11 

 species common to the Galapagos and Gulf of California regions and none 

 other, with the exception of the two anomalous species mentioned below, 

 have no Atlantic analogues. Nor have the 5 which remain common to the 

 Galapagos and Subantarctic region (Chile and Peru) alone, after the 

 exclusion of the four species previously mentioned of which the early 

 mainland records are in doubt. (Taliepus marginatus, a Peruvian species 

 not present in the Bay of Panama, is said to occur in the Atlantic in view 

 of the fact that its non-existent type was attributed, perhaps erroneously, 

 to Brazil.) 



The present study cannot be fairly concluded without mentioning two 

 species the recorded distribution of which cannot be satisfactorily ex- 

 plained by this reasoning. Leptodius cooksoni is found at Socorro and 

 Clarion, the Galapagos, and questionably, Chile. Furthermore, although 

 not present in the Bay of Panama, it is said to have an Atlantic analogue. 

 Any one of these relationships is possible, but the three together are in- 

 compatible with distributional theories which present the Galapagos as a 

 region of universal convergence only. Leptodius occidentalisj common to 

 the Gulf of California and Galapagos, but not found in the Bay of 

 Panama, is said to be analogous to L. floridanus. Here the species is 

 definitely of the northern fauna and the question becomes one of a past 

 link between the Gulf of California region and the Gulf of Mexico, 



