HYDROZOA CRASPEDOTA. 769 



this acrocyst protected by processes of the gonangium containing hollow extensions 

 of the tissues of the blastostyle, and together constituting a marsupium (Allman, 

 op. cit. pp. 50-3 ; Hincks, British Hydroid Zoophytes, p. 244). The Anthome- 

 dusan Eleutheria has a cavity aboral to the atrium and opening into the bell by 

 six interradial canals. The genital products are formed in the walls of the cavity, 

 which is also a brood-pouch (Hartlaub). So too in Pteronema and Clavatella x . 



Brooks has shown in the planulae of the Anthomedusan Turritopsis and the 

 Leptomedusan Eutima^ the presence of an ectodermal invagination, which is after- 

 wards evaginated and exudes a cement attaching the planula to some foreign body 

 (Mem. Boston Soc. Nat. Hist. iii. p. 402-3). 



A typical or monoprionid Graptolite, belonging to the section Graptoloidea, 

 consists of a hollow chitinoid coenosarcal tube or perisarc, which bears along one of 

 its aspects a row of tubular offsets, the cells or thecae. The tube is usually linear, 

 rarely leaf-like ; straight or curved ; simple or branched ; its walls composed of 

 three (?) layers. A solid or hollow chitinoid rod (=virgula) is imbedded in its walls 

 on the aspect opposite to the thecae. The proximal end of the coenosarc is connected 

 to a pointed dagger-like germ or sicula, from which it appears to originate, the pointed 

 end being imbedded in the coenosarc in some families, the broad in others. Two such 

 organisms may be united back to back by their virgular aspects, as in the diprionid 

 Graptolites, or even four, as in the Phyllograptidae. The virgula often projects in these 

 compound forms, more rarely in the monoprionid, far beyond the distal end of the 

 coenosarc ; and it may project slightly, as the radicle, at its proximal or sicular end. 

 Specimens of diprionid Graptolites have been found bearing large oval or triangular 

 capsules, supposed to be gonangia. Oval or bell-shaped capsules Dawsonia are 

 often found mingled with the Graptolites, and have been supposed to be detached 

 ovarial capsules (Nicholson). In the Retioloidea, the second section of Graptolithidae, 

 there is no sicula ; the outer layer of the perisarc is thin, the two inner reduced 

 to a network ; there are two opposite rows of thecae, but the coenosarcal tube is 

 common to both; the two virgulae have either coalesced, or are opposed to each other, 

 but in both instances situate in the perisarc. The colony is free in both sections. 



There is no indication as to what the animal of the Graptolite was like. All- 

 man has suggested that it resembled a Plumularian machopolype, inasmuch as the 

 thecae, like those of machopolypes, are not constricted at their bases as are those of 

 hydranths. This argument is not necessarily conclusive, for the thecae of Cuspi- 

 della (Hincks, op. cit. p. 209), or of Tricky dra (Id. p. 216, Fig. 26) are simple tubes. 

 The virgula is a structure unknown in any living Hydrozoan. The second stem, foun^l 

 by Jickeli in Anisicola (P lumularia] Halecioides, which bears no hydranths itself, but 

 supports the main stem to which it is connected at intervals, is hardly comparable 

 to it ; see M. J. viii. p. 636. Another difficulty is the fact that the virgula extends 

 far beyond the theca-bearing region of the coenosarc. Jickeli suggests (loc. cit. 

 pp. 668-71) that the Graptolites are possibly Octactinians. There are no known 

 Antipatharians with a chitinoid perisarc covering the outside of the organism ; but, 

 supposing there were, the virgula is not altogether unlike the chitinoid tubular 

 skeleton of Antipathes ; cf. p. 737. On the other hand no known Anthozoa possess 

 special enlarged generative capsules ; nor do the Polyzoa, a group to which th. 

 Graptolites have been sometimes assigned. 



1 Cf. note, p. 759, ante, and Haeckel, 'System,' pp. 101, 105. 



