PROF. ALLMAN ON STEPHANOSCYPHUS MIRABILIS. 63 



cavity they present a somewhat lobed or pouched appearance ; but beyond this they 

 become lost in the thickened walls of the body, so that their actual termination could 

 not be made out. Their walls are thick ; and their lumen, which is very narrow, is 

 doubtless lined by an endoderm ; but this could not be demonstrated with the same 

 distinctness as in the circular canal. 



The walls of the cavity which occupies the axis of the narrow portion, or stem, are 

 much thicker than those of the wide distal chamber, and posteriorly encroach so much 

 upon the axile cavity as to reduce this to a very narrow canal. 



Thread-cells exist throughout in the ectodermal layer ; but these are everywhere 

 exceedingly minute. Those examined from the body-walls in spirit specimens were 

 nearly spherical, and under a -^ immersion-lens were seen to include a spirally coiled 

 filament ; they were about the ^Vo °^ an ^ ncn * n diameter. 



The more posterior or proximal parts of the animal are fixed in the external chitinous 

 perisarcal tube, and possess no obvious contractility; but the more anterior or distal 

 portion is very contractile, and is quite free from the tube, so that it can move up and 

 down in it during extension and retraction, quite like a campanularian hyclranth in its 

 hydrotheca. 



The surrounding chitinous tube, indeed, increases so rapidly in width towards its 

 distal end as to resemble here the hydrotheca of a campanularian (fig. 3). It is marked 

 by annular rugae for some distance from its free extremity, and usually by fine longi- 

 tudinal striae ; but towards its attached end the marking disappears. 



At rather irregular distances along its length its inner surface sends off remarkable 

 processes, which project far into the interior (fig. 5, a, a, and fig. 3, e). At four equi- 

 distant points, situated on the same level, a thin chitinous lamina separates from the 

 inner surface of the tube, and forms a hollow pyramidal process. The four processes 

 thus formed radiate towards the axis, pushing before them the soft parts, and forming a 

 stelliform constriction of the body (fig. 6). Between each of these four radiating pro- 

 cesses a much smaller one is sometimes present, thus raising to eight the number of pro- 

 cesses in a set, and indicating an increase in number by consecutive intercalations of 

 younger ones between those previously formed. These imperfect septa, notwithstanding 

 their striking appearance, do not seem to possess any important significance in the orga- 

 nization. They will, however, suggest a comparison with the tabulse of the tabulate 

 corals, and may be regarded as offering some support to Agassiz's view of the hydrozoal 

 nature of these animals. 



In none of the specimens examined could generative elements be found. It is almost 

 certain, however, that the absence of obvious generative elements is the result of the 

 immature condition of the reproductive organs. "We should expect to find them in the 

 walls of the longitudinal canals ; and I believe that the poiiched condition observed in 

 these canals in the deeper part of their course is only an early state of the reproductive 

 system. 



