74 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 



their obvious differences have not been in real competition with kubiniji, 

 and which in the environment north of the equator, new to them, have 

 developed the numerous, slender abactinal spines which distinguish them 

 from their southern ally. The fact that microbrachius occurs at Panama 

 and Pearl Island may be interpreted to support this hypothesis. The 

 relationships of the island forms are obvious, for multiradiatus is very 

 closely allied to kubiniji, cumingii is quite as close to polybrachius and 

 canopus is almost certainly an offshoot from heiianthus. — These rela- 

 tionships, both phylogenetic and geographical, may be indicated by such 

 a sketch as Diagram 6, it being understood that the relative length of 

 the lines has no significance whatever. 



Because of the extremely littoral habits of Heliaster, there can be no 

 question that the island forms have reached their present homes as 

 larvae transported by ocean currents. Owing to the distances however 

 and the slow rate of travel, the chance of survival is very small, and it 

 must be seldom indeed that young Heliasters from the mainland ever 

 reach the Galapagos or even Juan Fernandez. The latter islands seem to 

 have been reached as yet only by the single species (heiianthus) from the 

 nearest mainland, which under the stress of new conditions has become 

 changed so that it breeds earlier in life, and is consequently much smaller 

 than its parent form, and has more delicate spines, and fewer, freer rays. 

 The Galapagos have been reached by young polybrachius from South 

 America and also by young kubiniji from Mexico, but if we may judge 

 by the relative amount of change, Juan Fernandez was populated by 

 Heliaster long before the Galapagos. At the latter islands, cumingii 

 appears to be much more abundant than multiradiatus, so we are justi- 

 fied in thinking polybrachius was the first comer, but both are so recent, 

 the changes are as yet slight. 



Of the factors which have led to this development of diverse forms of 

 Heliaster, one at least stands out so clearly that there can be little doubt 

 of its importance, and that is isolation. Were only the maiidand species 

 known, this factor would not be so obvious, though it would be suggested 

 by the apparent lack of Heliasters on the coast of Colombia. But when 

 we consider the two Galapagos species, and particularly when we study 

 canopus, it is hard to doubt that the complete isolation of these small 

 groups of individuals has been of great importance in the formation of 

 the new species. In the case of canopus, there has been sufficient time, 

 so that the species is sharply distinct, while the Galapagos species seem 

 to be as yet only imperfectly defined. It is not necessary to claim that 

 isolation has been the only, or even the essential, factor. Indeed the 



