304 DISCOVERY REPORTS 



The Falkland species apparently passes through a series of stages in development 

 somewhat similar to those gone through by the British species Willia stellata. A brief 

 description of the stages of W. stellata may help towards the elucidation of the stages 

 of the Falkland species which is subject to so much variation. The development of the 

 medusa, W. stellata, may conveniently be separated into four stages which are based 

 upon the radial canal system: 



First stage: Stomach with six lobes, from each of which runs a radial canal without 

 any branches (called the main radial canal in the later stages) to the margin of the 

 umbrella. Six tentacles, one opposite each canal. 



Second stage : Six main radial canals, each with one lateral branch. Twelve tentacles, 

 one opposite each canal. 



Third stage: Six main canals, each with two branches, opposite each other, all 

 running to the margin. Eighteen tentacles. 



Fourth stage: Six main canals, each with three branches; the third branch comes 

 from a lateral canal and not from the main canal itself. Twenty-four tentacles. (Adult 

 stage.) 



It will be seen that as the medusa grows the radial canals become branched; the 

 branches increase in number, and a corresponding increase takes place in the number of 

 tentacles. 



The mode of formation of the branches in W. ttiutabilis is somewhat uncertain. 

 Blindly ending canals are very rarely seen, but when a young tentacle bulb is just 

 beginning to make its appearance on the umbrella margin, it is usually seen to be con- 

 nected with the previously existing canals by a delicate, almost imperceptible streak of 

 cells. Apparently, therefore, each new branch is the result of a differentiation of cells 

 in the endoderm lamella of the umbrella taking place in the whole course of the future 

 canal branch simultaneously, and not progressing from one end to the other. 



Each new tentacle makes its appearance below one of the exumbrellar clusters of 

 nematocysts and for some time, during the growth of the tentacle, the nematocyst cluster 

 is seen above the young tentacle, but later on it is gradually displaced to one side. 



As pointed out by Browne (1896, p. 471, and 1897, p. 823), abnormal specimens of 

 W. stellata may occur, having a number of main radial canals other than six. In the 

 Falkland specimens such irregularities are much more frequently found. For the 

 description of the development and the variations in W. mutabilis it will be convenient 

 to divide the specimens into three groups: (i) Specimens with eight main radial canals 

 regularly arranged. (2) Specimens with six main radial canals regularly arranged. 

 (3) Irregular specimens. Within each of these groups the specimens are subject to some 

 individual variation ; specimens in the same stage of development of the canal system 

 are of varying size, and the branching of the canals does not always proceed at the same 

 rate in all of the main canals in one and the same specimen. Apart from such individual 

 variations the development is as follows: 



(i) Specimens with eight main radial canals, regularly arranged (forty-one specimens 

 examined). In the youngest specimens observed, about i mm. in width, the stomach is 



