REPORT ON THE TUNTCATA. 123 



account of their connection with ancestral Chordata, (2) because of the well-marked 

 degeneration which they exhibit, and (3) Ijecause of the curious course which the 

 group has apparently followed in its evolution, one result of which is that one of the 

 largest and most important sections, that of the Compound Ascidians, is, as I shall show 

 farther on,' polyphyletic in origin, and is consequently an unnatural group. 



Assuming the protochordate relations of the group, our conception of the ancestral 

 form ' from which the primitive Tunicata were derived would take the form of an 

 elongated 1 )ilaterally-symmetrical free-swimming animal with a metamerically segmented 

 body, terminating anteriorly in a prseoral lobe. The mouth would be ventral and the 

 anus posterior, and the anterior portion of the alimentary canal would be provided 

 with a series of laterally-placed respiratory slits putting paired diverticula of the cavity 

 of the fore-gut in communication with the exterior of the body. On the dorsal 

 surface of the alimentary canal would Ije placed a median hypoblastic rod, the 

 notochord, and above that again the dorsally-placed nervous system, possibly still in 

 connection with the epiblast over it, and forming a longitudinal median tract 

 probably produced by the union of two lateral nerve cords, and provided with an 

 anterior enlargement in the pra3oral lobe, and possibly with a ganglionic thickening in 

 each metamere. The ccelom, formed by the union of outgrowths from the archenteron, 

 would have paired nephridia of the typical form, placing it in connection with the 

 exterior of the body, and probably one or more such pairs were present in the prseoral 

 lobe in close relation with the ventral surface of the large nerve mass. 



From such a form as this the primitive Tunicata might be evolved with a 

 slight amount of degeneration. Nearly all the existing groups of Tunicata pass 

 through a free-swimming larval stage, which probably represents very closely the 

 structure of a common ancestor not far removed from the existing Aiypendicularia. 

 Such a hypothetical form would differ mainly from its protochordate ancestors in 

 having the notochord limited to the posterior part of the body, and not extending 

 forwards into the region occupied hj the chief parts of the nervous system and 

 alimentary canal ; and in having the anus ventral in place of posterior. It may 

 readily be imagined that some group of the free-swimming Protochordates would find it 

 an advantageous modification that the alimentary canal, which performed both nutritive 

 and respiratory functions, and the main part of the nervous system, in connection witli 

 which sense organs had Ijecome developed, should be as much as possible concentrated in 

 the anterior part of the body so as to leave the posterior part free to l)ecome modified 

 into an efficient locomotory organ. Under such circumstances it would be natural that 

 the notochord, the sole internal skeleton, should become restricted to the posterior, 



1 See also this Keport, Part II. p. 387. 



- See Herdman, Phylogen. Classif. of Animals, p. 57, 18S5 ; and Van BeneJen and Julin, Morpli. d. Tuniciers, 

 p. 415, 1887. 



