306 Heredity. 



whichever one we accept, we are compelled to believe 

 that there has been parallel evolution, and that certain 

 homologies between the various forms are not due to in- 

 heritance, but to independent modification. 



Haeckel's view compels us to believe that the resem- 

 blances between the two groups of medusae have been 

 independently acquired. If we accept the second alter- 

 native, and derive the two forms from an ancestral me- 

 dusae, the resemblances between the larvae must be due 

 to parallel variation. Suppose, then, that we accept the 

 third or the fourth view, and derive both groups from 

 an ancestral form which had a polyp-like larval stage, 

 and a medusa-like adult stage, or else from an ancestral 

 form which united in itself certain of the larval and cer- 

 tain of the adult characters. 



Among the veiled medusae there are some which, like 

 Liriope, do not pass through a hydra stage, but lay eggs, 

 which develop directly into medusae; and there are 

 also forms which, like the fresh water hydra, have no 

 medusae stage. Among the veilless forms there are 

 also some which have no medusa stage, but which, like 

 Lucernaria and the Tesseridae, remain permanently as 

 scyphostoma-polyps, and it is probable that in others, as 

 the Charybdeadae, the eggs hatch into medusae, as they 

 do in Liriope, without the intervention of a larval polyp 

 stage. 



It is therefore impossible to frame any hypothesis as 

 to the origin of the medusae, which will do away with 

 the necessity for the belief that parallel modification, 

 along independent lines, has occurred in the different 

 subdivisions of the group. 



If we accept Darwin's view of the origin of varia- 

 tion, parallel modification is not absolutely impossible, 

 although the chances against it are very great indeed. 



