TUNICATA OF THE PRIBILOF ISLANDS. 523 



years. It ha^ several representatives ou the coast of Califoruia, so I have had ample 

 opportuuity for stiulyiiig it; uotwithstaudiug this, I have not been able to fully satisfy 

 myself as to whether a new genus should be established for it or not. That it belongs 

 to the family Polyclinidae there can bj no doubt. The possession by the zooids of a 

 large, well-marked post-abdomen in which are situated the reproductive organs and 

 heart leaves no room for question on this point. When, however, the effort is made to 

 determine witli which of the known genera of this family the group is most closely 

 allied, much diflBculty is experienced. 



The entire absence of vsystems or coenobia in the colonies leads us, in the first 

 place, to compare it with those Polyclinidae presenting a like deviation from the pre- 

 vailing condition in this particular. 



In the genus Tylobranchion, Herdman, no common cloacal apertures are present, 

 according to the author, but this is about the only resemblance between the two 

 groups; the most distinctive difference being, perhaps, the possession of papillae ou 

 the internal transverse vessels of the branchial sac in Tylobranchion. 8i(/illina, 

 iSavigny, is another genus in which the common cloaca is wanting; but the shortness 

 of the branchial sac and great length and slenderness of the i^ost abdomen are char- 

 acters which preclude the admission of our species to this group. Sigillina australis, 

 Savigny, the onl}^ species known of this genus, has but four series of stigmata, while 

 there are never less than six or seven present in any of the rei)reseutatives of the ' 

 groui) now under consideration, and the rule is that twelve or thirteen series are 

 present. 



As regards the post-abdomen in Sigillina, its great length, relative to the length 

 of the rest of the animal, and its tenuity, set it off very sharply not only from our 

 forms, but also from all other known species of the family. 



The genus Atopogaster, Herdman, contains one species, at least, viz, A. aurantiaca, 

 in which, according to this author, there are no systems or common cloacal apertures, 

 and there are certainly some rather weighty considerations in favor of regarding this 

 group as the one to which the present species is most closely related. Several points, 

 however, the most important being the transverse folds of the stomach wall in Atopo- 

 gaster, stand rather seriously in the way of doing this. There is considerable varia- 

 bility in the character of the stomach wall in our species, and the folds are never well 

 pronounced. Such as are present, however, incline distinctly toward the lengthwise 

 instead of toward the crosswise direction of the stomach. 



I am disposed to place somewhat less reliance than some writers have done on 

 this character as an index to relationships; nevertheless a condition so unusual as 

 a transverse folding must, as our knowledge now stands, be regarded as of real 

 systematic value. 



I have resolved, after much deliberation, to place the species, for the present at 

 least, in the genus Aplidiojysis, Lahille. There are certainly some objections to this, 

 the most considerable being found again in the structure of the stomach wall. Lahille 

 instituted this genus for the reception of those Polyclinidae in the restricted sense in 

 which he recognizes this family, which have a smooth walled stomach, no torsion of 

 the intestinal loop, and a nonpedunculated post abdomen. The smooth wall of the 

 stomach, therefore, is one of the important characters on which the genus rests, and 

 the placing of my species in it does some violence to it, for there is certainly a strong 

 tendency, to say the least, for the stomach wall here to become folded, i. e., there are 



