I0 STYLASTERIDAE 



When the material available only consists of some few fragments of Stylaster gemmasccns, we 

 may often be inclined to divide it up into several species, as already mentioned. Some pieces 

 (PL I figs. 5 6) show a regular aspect with uniformly constructed branches; others on the other hand 

 (PI. I, figs. 4 and 7) are irregular in their mode of branching and in the arrangement of the cyclosystems. 

 A larger material soon shows however, that there is no justification in creating subspecies on the 

 basis of these variations. Often the appearance of the ampullae may alter the regular structure of the 

 colony, though not quite so much as in the following species, with which it can easily be confounded. 

 The most reliable mark of distinction between the two species is found in their cyclosystems and 

 gasterostyles. But the female ampullae also afford good characters; in Stylaster gemmascens they are 

 equipped with a varying number of conical, blunt spines; normally we find 3 or 4 of these on the 

 ampulla, more seldom 2 or i ; often the number may also be larger and once 7 spines were found on 

 a single ampulla. When the ampullae occur in larger number, these small spines appear in quantities 

 between the cyclocystems and make the boundaries and arrangement of these indistinct to the naked 

 eye (PI. I fig. 7). Closer examination however shows that the cyclocystems are comparatively little 

 affected by the ampullae. 



The cyclosystems appear as small, oval, stellate elevations except at the very tip of the branches, 

 where they are still circular in circumference. Even the tops of the septa-like separating walls between 

 the dactylopores project somewhat strongly above the surface of the colony. The number of the dac- 

 tylopores in the cyclosystems varies from 12 to 20; in general there are from 14 to 18 dactylopores 

 round a gasteropore. The wall of the gasteropore shows a deep incision towards the dactylopore. 

 Any fusion of the dactylopores which lie side by side could not be detected, as is often the case in 

 some tropical species. - - Carefully prepared longitudinal sections made by grinding readily show the 

 gasterostyle (PI. Ill fig. 21); in Stylaster gemmascens it is pointed, conical and almost twice as high as 

 broad. A thin section (PI. Ill fig. 24) shows clearly its lattice-work structure. The dactylostyles are reduced 

 to faint, irregular elevations on the outer wall of the pores; they are very difficult to observe. - 

 Thin sections of the stem and main branches show very distinctly, that the growth here as in our other 

 northern Stylaster species proceeds centrifu gaily and periodically; but it still remains to discover, what 

 influences in the sea produce this periodic growth in the colonies. 



The gasterozooid has the same structure as the hydroid polyp, when we exclude the secondarily 

 formed gasterostyle (Text-fig. C, PI. Ill fig. 25). Partly near to and partly somewhat below the opening 

 of the mouth we find four quite small, almost rudimentary tentacles. The small cnidocysts of the 

 polyp are gathered in these tentacles, though without leading to the formation of a typical, thickened, 

 distal part such as is found on the capitate tentacles of the Corynidae\ their structure agrees fully with 

 the thread-like tentacles we find in the athecate Hydroids. The ectoderm of the gasterozooid is nor- 

 mally destitute of stinging cells otherwise. The cell boundaries in the ectoderm are very difficult to 

 make out; the whole structure here agrees with that of the Hydroids. Often we can observe out- 

 runners which connect the ectoderm of the polyp wall with the epithelium of the gasteropore wall; 

 these resemble the irregular outrunners, which are often found in the thecaphore Hydroids and which 

 here connect the ectoderm of the hydranth with the hydrotheca outside the true, basal line of attach 



