PAEDOPHORE HYPOTHESIS 359 



At first, no doubt, the actinulae swam by means of tentacles, as suggested by Garstang (1946), 

 before settling down on the substratum and budding-off the sexually mature medusae. Soon asexual 

 reproduction or budding took place during the prolonged, free-swimming larval phase, and finally the 

 sexual adult stage was budded from the daughter polyps (gonozooids) of the still free-swimming 

 actinula, originally a larva but now a neotenic, mature animal of a new type. So now the 

 parent zygote (oozoid) has become an asexual carrier of the offspring (gonozooids) that I call a 

 paedophore. 



I do not propose to try to suggest all the detailed steps that must have followed one another in the 

 course of the origin and adaptive radiation of this new type of organism, the paedophore. But some 

 of the animals evidently remained entirely free-swimming and emerged as disconanths, Pelagohydra 

 and some of the margelopsines, while others took to the bottom and emerged as loosely attached 

 myriotheline and corymorphine hydroids. The great interest of Marge lopsis haeckeli (Werner, 1954) 

 is that it is planktonic for most of the year. During this time the larvae are retained until they 

 develop into actinulae, which are then set free to become free-swimming hydroids ; but the autumnal 

 eggs develop into stereoblastulae, which pass the winter on the sea-bed. This life-cycle indicates how 

 an entirely free-swimming hydroid may have arisen from a fixed precursor as a neotenous larva. 



The essential structural organization of the new type of actinula is the possession of two sets of larval 

 tentacles, oral and aboral, together with daughter polyps in the region between. These daughter 

 polyps either have functional mouths and are called gonozooids, as in the disconanths, or the mouths 

 are lacking although the polyps have tentacles and are obviously reduced polyps, as in Myriothela 

 penola Manton. Again, the daughter polyps may be so reduced as to form scarcely recognizable stalks 

 which bear the medusae, as in Margelopsis haeckeli Hartlaub. 



In siphonophores generally, these reduced gonozooids are difficult to identify because often 

 there is a replication accompanied by successive reduction. This can be seen in Physalia, where each 

 gastrozooid except the protozooid may be regarded as the first of a series of groups (cormidia) of 

 gonozooids, the last gonophores and nectophores being produced on the last few of the series of 

 replicates called palpons. 



Garstang agreed with Schneider who, in 1896, expressed the view that the siphonophore somatocyst 

 was homologous with a bracteal canal. These canals Garstang showed to be homologues of tentacles. 

 Leloup in 1954 suggested that a siphonophore tentacle was homologous with an oral tentacle of an 

 actinula. Neither of these zoologists pursued the subject to the conclusion at which I myself have 

 arrived. In fact Garstang still thought that the nectophore of a siphonophore represented the cauline 

 gonophores of a stalked hydroid, and not those of hydroids like the corymorphines that bear them on 

 the hydranth. In my view it is with the gonophores of the hydranths that the nectophores are to be 

 homologized. We can in fact put the figure of a swimming margelopsine hydroid, Climacocodon or 

 M. gibbesi, which is really a neotenous actinuloid nurse-carrier, side by side with that of either a 

 calyconula or a physonect larva of a siphonophore and show correspondence of essentials (Text-fig.30). 

 Although Climacocodon has radial symmetry and many tentacles arranged in two groups, oral and 

 aboral, and produces sexual medusae, while the calyconula has only one of each group of tentacles and 

 a single asexual medusa arranged on a single meridian, Garstang has provided an explanation for this 

 difference in symmetry. Bilateral symmetry in the calyconula was due, he pointed out, to the presence 

 of a great deal of yolk which restricted the delamination of tissues to a single meridian. The calyconula, 

 no doubt, is not very like the larva of the ancestral siphonophore, which probably had a float as does 

 the larva of the cystonects, the group which includes Physalia. Garstang showed how the calyco- 

 phores had lost this float. Like the larva of the physonects, the calyconula has a precocious necto- 

 phore, which the cystonula larva of Physalia has not. 



8-2 



