( 845 ) 



also have possessed unpaired tentacles and an unpaired gastral division 

 in the median line. They forget that such an unpaired medial coeloraic 

 cavity is already present in Balanoglossus and that Langerhans 

 ("Zeitschr. fiir wiss. Zool." ^'ol. 34. 1880) and (toodrich (Q. J. 

 Microsc. Sc. Vol. 44, 1901) have also shown the existence of an 

 unpaired coelomic cavity in Saccocirrus, while cases of unpaired 

 median sensory spots could be enumerated in Coelomata. The well- 

 known antithesis of headseginent and pygidium as compared with 

 the trunk in the bilaterally metamei-ic Coelomata is an argument 

 that goes far to meet Lang's and Hatschek's contention. 



Neoformation of segments, still pretty equally distributed with the 

 cyclomeric Coelenterates, but there also already \ariable, occurs with 

 the Coelomata exclusively at the posterior end and with many of 

 them only during embryonic life. 



If we assume in accordance with E. van Hkneden (see Verh. Kon. 

 x4kad. V. Wet. Amsterdam. Vol. VIII p. 69) that the stomodaeum 

 of an actinia-like ancestral form has been the precursor of the chorda 

 dorsalis, beside and above which the ner\-e-rijig unites to a spinal 

 chord, while under it a connection between the stomodaeal fissure 

 (the chordal cavity) and the gastral diverticula (coelomic cavities) 

 can be observed, then it follows from this that the ancestral forms 

 of the acpiatic Chordata are moved about in the water in a position 

 diii'erent from that of the ancestral x-Vrticulates by 180'. The mouth 

 of the Chordates would then have arisen later as a new formation, 

 as has already repeatedly been asserted. It is an undoubted simpli- 

 fication of our phylogenetic speculations if we are entitled to look 

 for this difference in orientation already in \ery early ancestral 

 forms and can so avoid Geoffroy St. Hilaire's error who shifted 

 the process of reversion to a much later stage of development. 



If thus the phylogenesis is very considerably shortened, I may 

 call attention to the tact that even with respect to the mammalian 

 foetal envelopes, I showed in the al)Ove-mentioned publication the 

 possibility of a similar shortening of their phylogenesis, by not 

 admitting that these embryonic coverings have originated from those of 

 reptiles and birds, as was done until now, but by considering them in 

 direct connection with larval envelopes of invertebrate ancestral forms. 



In the grouping of bilaterally symmetrical, coelomatic animals 

 (resp. of such as have lost their coelom again) which has here 

 been attempted, important phyla (Xemertea, Nematoids' Prosopygii, 

 Chaetognatha, etc.) have been left out of consideration. 



There are still too many gaps in our knowledge of their anatomy 



