240 A. K. TOTTON 



bases of the first two gastrozooids, the area which bears these bracts takes up a 

 subterminal position at the end of a long stem. These larval bracts, so often figured, 

 and their successors of similar though slightly modified shape, are quite distinctive; 

 and differ from the definitive bracts, which arise in hundreds proximally to them in 

 several ventro-lateral meridians of the stem. The contrast in bracteal types is best 

 seen in early stages of metamorphosis, when the definitive bracts first make their 

 appearance. 



Information has been wanting in published reports on the position and time of 

 first appearance of the nectophores. Garstang suggested that, as the nectophores 

 appeared, so the bracts, which hitherto had served for locomotion, dropped off, 

 but as we have seen this is not so. The oldest of Metchnikoff's figured specimens, 

 in what might be called the Me/ophysa stage of development, shows two functional 

 nectophores, but unless the reader is familiar with the animals it is not easy to orientate 

 and interpret the figure. The little group of four rounded buds, seen through the 

 uppermost bract, probably represent buds of gastrozooids and palpons. Close by, 

 though not figured, is the nectostyle to which the larval bracts are attached by their 

 muscular lamellae. The nectostyle can be seen in sagittal section (Fig. 1a) to be an 

 elongated, hollow, cone-shaped diverticulum of the general cavity. The larval bracts 

 are attached by muscular lamellae at diflFerent levels on many meridians on either side 

 of the ventral line from which the cormidia are budded. From the opposite side of the 

 pneumatophore spring the nectophores. As pointed out by Garstang (1946), 

 delamination of tissues starts on one side (called ventral, on which the nectostyle is 

 found), and proceeds gradually towards the opposite, dorsal side. This explains why 

 the nectophores are late in appearing where they do on the dorsal side. Metchnikoff 

 says that the pneumatophore has become free, in his figure 1, through the loss of one 

 of the larval bracts. He draws attention to the beginnings of a stem; to the one-sided 

 position of the nectophores; to the first definitive bract below them; to four palpons; 

 and to the two (larval and first definitive) tentacles and a single gastrozooid. Evidently 

 he was unaware of the smaller protozooid or confused it with a palpon. This is the 

 only published account known to me of the beginnings of metamorphosis in a physonect 

 siphonophore. 



But curiously enough metamorphosing Physonects, probably of more than one 

 ■species, have been figured unwittingly as adult representatives of a distinct genus 

 " Nectalia ". Garstang (1946, pp. 172-5) spent much time discussing " Nectalia " 

 He quite correctly perceived that its long bracts were precormidial, or coronal as I 

 prefer to call them, and homologous with the bracts of Athorybia rosacea and of 

 Melophysa melo. But the reasons for his conclusion appear to be wrong in that he 

 implied that these bracts would not be carried downwards with increase in number of 

 the cormidia. The coronal (larval) bracts of Agalmids are in fact carried downwards 

 with the first and second gastrozooids. They are " precormidial " in the earliest 

 stages only. New bracts of this coronal type are formed for a long time after meta- 

 morphosis, but only in this restricted area, which comes to be terminal, and on the 

 distal (oral) side of their predecessors. The later-formed bracts of this type are much 

 longer than the first-formed ones, and gradually diverge from them somewhat in 

 shape. But from the start they have a distal pocket of nematocysts at the end of the 

 bracteal canal (misinterpreted by Haeckel as a medusoid subumbrella), whereas 

 the cormidial (definitive stem) bracts in Agalma elegans do not. 



