99 



the connection being visible in sketch VIII (<?). In this same sketch \'III we find that these three 

 cavities are gradually diminishing in size, the most proximal one being already closed and the 

 resemblance with the secondary ectoderm oi Athorybia and such a cavity formed by this ectoderm 

 as figured by Chun (97 a on Plate IV, fig. 7) is a very striking one. 



Finally the last sketch IX (PI. XVII, fig. 133 IX) shows how the cavities have partly 

 disappeared, the most distal one being the only one left. In sketches VI — IX we have seen 

 an increase of the developing chitinous substance in the cushion-like secondary ectoderm. This 

 substance corresponds perhaps to the darker green colour situated between the two chitinous 

 enlargements. 



As we said already above, we tried to get a more definite idea of the structure of the 

 appendages, but by reason of their histological insufficiency we had to give up this examin- 

 ation entirely. 



The sketch of the mature appendages in the diagram of the whole, may perhaps seem' 

 deceptive, as on the left near the air-funnel the bigger siphons are drawn, and on the opposite 

 side the smaller ones. It would seem therefore that they develop on the side opposite the one 

 we found to be the zone of proliferation. To explain this apparent mistake we may add 

 that an enormous quantity of appendages (which would have filled up at least one more plate 

 underneath the diagram (PI. XVII, fig. 124) could not be added. That the siphon on the left 

 side has probably been cut right through the middle, whilst the others on the right are only 

 touched superficially which accounts also for the appearance of the two ovate bodies on the basal 

 side. The two nectophores have not been included in the diagram, the}- develop themselves in 

 the outer wall, about the same height as the furthest appendage [a.s.a^j shown on the diagram. 



After the description of the internal structure of Archangelopsis ^ we think it no longer 

 difficult to remould Haeckel's Aiironectae in such a wav as to make the homoloev between 

 zone of proliferation and air-funnel on one side and aurophore on the other a striking one. 



If we compare Haeckel's longitudinal section of the aurophore of Rhodalia on PI. V, 

 fig. 24 of his Challenger Report with the series of sections given on PI. XVII of this work, 

 we find that by diminishing the fulcrum ("::" on PL \^ Haeckel 88b), the ectoderm follows the 

 folds of the entoderm on the other wall ; they become the youngest buds of appendages. The 

 pericystic cavity in Haeckel's Rhodalia continues (the porus being accidental) and the entoderm 

 on the opposite side constitutes one of the layers of the pneumatosaccus. 



The irregular folds correspond to the buds of youngest appendages. 



The outer ectodermal layer also runs smoothly along the "spherical bag-like structures" 

 in Fewkes' Angelopsis globosa. Fewkes has already seen in the cavity of one of them "bodies 

 resembling those of the lower floor" i. e. appendages of the siphosome (84 p. 973). This means 

 that Fewkes suspected that the internal structure was more complicated than he could at that 

 time conceive. We believe that both Haeckel's Auronectae and Fewkes' Angelopsis globosa 

 were in a different state of contraction from ours, and that the latter represent by far the most 

 natural condition. 



