THE Ancestry of the vertebrates 631 



up indifferently, and that dorsal and ventral sides may be reversed in 

 relation to the earth in different genera of the same phylum. Among 

 Crustacea for example many phyllopods swim with the neural side upper- 

 most. This is the normal position of most adult cirripeds. Amphioxus 

 in swimming maintains no fixed orientation of its body in relation to 

 gravitation. 



The main supports of the annelid hypothesis are found in metamerism, 

 terminal growth, and the circulatory and urogenital systems. Both 

 annelids and vertebrates have dorsal and ventral longitudinal blood 

 vessels, interconnected around the pharynx by aortic arches. If the 

 dorsal and ventral sides of the annelid are reversed, the flow of blood is 

 similar, and the contractile dorsal vessel of the worm becomes the heart 

 of the vertebrate. The heart of amphioxus and that of vertebrate 

 embryos is, as would be expected from the annelid theory, a median 

 longitudinal tube. In both phyla, similar visceral and somatic vessels 

 occur. The cardinal veins of chordates, however, are wanting in annelids. 



Semper, in supporting the annelid theory, laid great stress upon the 

 similarity of segmental tubules in annelids and chordates. The pro- 

 tonephridia of amphioxus are identical with those of annelids, and the 

 coelomoducts of annelids resemble the renal tubules of vertebrates 

 both in development and in metameric relations. Some annelids have 

 both nephridia and coelomoducts in the same metamere. Apparently, 

 nephridia have persisted in cephalochordates and coelomoducts in verte- 

 brates. The failure of earlier morphologists to recognize the difference 

 between nephridia and coelomoducts should, therefore, not prejudice 

 the case against the annelid theory. 



The presence of a longitudinal collecting duct, into which the nephridia 

 open in such annelids as allolobophora, has been cited as evidence in 

 support of the annelid theory. Superficially this collecting duct might 

 seem to resemble the primitive duct of vertebrates. But the resemblance 

 is misleading, since the collecting duct of annelids is ectodermal in origin, 

 while the primitive duct of vertebrates is mesodermal. The two ducts, 

 therefore, cannot be homologous. The evidence, instead of supporting 

 the annelid theory, actually weakens it, since it shows that structures not 

 genetically related may closely resemble one another. In other words, 

 this evidence appears to favor the alternative theory that the resemblances 

 between annelids and chordates illustrate the principle of convergence. 

 If the ectoderm of annelids and the mesoderm of vertebrates can produce 

 organs so similar in form and function, we are impressed more by the 

 plasticity of organisms than by their genetic relationship. 



On the other hand, the gonads of annelids and chordates appear to be 

 comparable. But there are no facts which invalidate the assumption 

 that the gonads of annelids and chordates have evolved independently 



