I02 



we found, however, these beak-Hke tentilla only in the very oldest specimen (Cat. 18), whilst all 

 the other specimens showed the trifid type and the palmate. So judging from our material we 

 conclude that the beak-like are not the very oldest tentilla, those which develop first on the 

 stem. But we cannot come to any definite conclusion, as after all preserved material of Sipho- 

 nophores is the very worst for deciding the question of larval stages. We simply describe the 

 facts, leaving it to future investigators, who are fortunate enough to examine fresh material, to 

 decide these important questions. 



Between the ninth and tenth, and the tenth and eleventh siphons two gonodendra are 

 developed which have already the characteristic shape and are easily recognized as such. 

 Younger gonodendra than these two could not be distinguished from the other buds. 



Cat. 18 (PI. XVIII, fig. 143 — 145, PI. XXI, figg. 151, 152) was very much entangled 

 when we found it amongst the Siboga material. The middle-part of the stem had absolutely 

 wound round the pro.ximal part near the pneumatophore. Moreover tentacles, tentilla and siphons 

 had clung so tightly around the twisted double stem that it was only with the greatest difficulty 

 that we succeeded in disentangling it all, at the cost of a few tentacles and siphons. But we 

 managed after all to get a complete stem, in which the appendages show their development 

 at regular intervals. It has now a length of 63 cm. 



The pneumatophore (long g mm. broad 6 mm.) shows no irregular shape, it has the 

 same structure as Gegenbaur describes for Rhizophysa filiformis. 



The stem narrows suddenly immediately below the ])neumatophore; it has then an 

 absolutely thread-like appearance which it keeps all along but for some irregularities in its distal 

 part. Near the pneumatophore we find a cluster of buds of siphons and five young siphons. 

 Then follows a short space where already the gonodendron can be distinguished as a very 

 young spherical swelling on the stem. Remains of tentacles also are still situated on that part. 



The further part of Cat. 18 shows 22 siphons which have partly lost their tentacles 

 (e. g. the third, the fourth, etc.). A small stump at the base of these siphons shows, however, 

 where they have been attached to the stem. The tentilla are well developed ; we found again 

 the three types Gegenbaur originally observed. EspecialK' the trifid type is well preserved 

 (PI. XV^III, figg. 143, 144); the palmate tentillum has undergone much alteration but still its 

 shape is easily recognizable. The appearance of irregular hard white granules makes it easy to 

 find these tentilla on the stem. A\'hat the significance of these hard (? calcareous) granules is, 

 has not yet been discussed. 



The third type (PI. XVHI, fig. 145) can with a little difficulty be identified with Gegenbal'r's 

 figure (54 PI. XVIII, fig. 9). The beak instead of standing straight out is bent to the proximal 

 part of the tentillum, its point being covered over by the numerous threads of exploded nema- 

 tocysts which arise on the ventral part. 



We found only a few of this last type and we cannot say anything definite as to the 

 exact position of the tentilla on the stem, as we often found detached tentacles, even detached 

 tentilla on the naked parts of the stem. Gonodendra are present in every .stage of development. 

 Between the ninth and the tenth mature siphon we find the most developed one, after the 

 eleventh only the basal part of the stalk (gono.styIe) is left, the gonodendron itself being detached. 



