64 VERTEBRATE ZOOLOGY 



The Pharyngeal Clefts.— These structures take the form of "gill- 

 sacs," each of which opens into the pharnyx by a U-shaped slit, re- 

 sembling that of Amphioxus, and opens to the exterior by a small 

 pore. These "gill-sUt" openings through the pharynx are sup- 

 ported by thin chitinous iars, also resembling the gUl-bar system of 

 Amphioxus. 



The Neural Tube. — The nervous system of Balanoglossus is in gen- 

 eral quite unlike that of the true chordate. Though it is diffuse and 

 rather poorly centrahzed, a distinct ventral as well as a dorsal nerve 

 cord occurs, the two being connected at the base of the collar by a com- 

 missure. In the collar region a short posterior part of the dorsal nerve 

 cord seems to be distinctly tubular. . In Glossobalanus and Ptychodera, 

 two other enteropneustan genera, the entire dorsal nerve cord of the 

 collar is said to be tubular. The claim therefore of the Enteropneu- 

 sta to chordate affinities are on this score rather strong, for no other 

 phylum of animals has a tubular dorsal nervous system. 



The most pronounced dissimilarities to the Chordates are seen in 

 connection with the coelomic cavities, for there are five separate cav- 

 ities: an unpaired proboscis cavity, paired collar cavities, and paired 

 trunk cavities. In this respect there are rather striking suggestions 

 of echinoderm conditions. 



The Tornaria Larva. — Perhaps the most marked evidences of echin- 

 oderm affinities, however, are seen in the larva of Balanoglossus, the 

 Tornaria larva (Fig. 46) which is so strikingly Hke that of the young 

 Auricularia larva of a holothurian that it was originally classed as an 

 echiaoderm larva by Johannes Miiller. The relationship that seems 

 to be indicated is not so much between the Enteropneusta and the 

 modern echinoderms as between a remote bUateral ancestor of the 

 echinoderms and an equally remote pelagic ancestor of the Enterop- 

 neusta. It is probable, however, that the rather superficial resem- 

 blance between the larvae of these two groups has been overempha- 

 sized. Larval resemblances present at best a very uncertain basis for 

 phylogenetic conclusions, since so many of the structures present are 

 mere provisional larval organs of probably caenogenetic origin. 



We must then conclude that the Enteropneusta show evidences 

 of having been derived from a stock that was related to the remotest 

 ancestors of the chordates (the so-called protochordates) and to-day 

 represent a rather unsuccessful lateral offshoot from the base of the 

 chordate branch of the phylogenetic tree. The closest linkage be- 



