STVI,ASTERIDAE II 



ment. — The gasterozooid is provided with a strongly developed, supporting lamella. The endoderm 

 cells are high and show the same cell types that are usually found in the lower Coelenterata; the en- 

 doderm continues on to the gasterostyle and forms its epithelial covering. Distinct cell boundaries 

 can in general not be observed in the endoderm of the gasterostyle; the granulated 

 structure shows, that the albumen cells are decidedly in majority here, whilst 

 Schneider's "Nahrzellen" (1902 p. 579) compose the main mass of the cells in 

 the endoderm of the free gasterozooid wall. 



A transverse section of the basal part of the gasterozooid gives an extremely 

 characteristic picture (PL IV fig. 32). The tissues collect into separate columns 

 which gradually become smaller in diameter and pass over into the stolons. No 

 lumen could be detected in these columns until they change over into the typical 

 stolons. Nor do they show any sign of differentiation into ectoderm and en- 

 doderm like the stolons. The number of columns does not seem to be constant. 



The dactylozooids (PI. IV fig. 33) have a very thick and muscular sup- Text-fig. c Diagram- 



. , . matic median section 



porting lamella and show a more marked power of expansion and contraction through the cvclosvstem 



than the gasterozooid. The thick ectoderm of the free wall of the dactylozooid of stylaster (Emtyiaster) 



. gemmascens. The tentac- 



is densely beset with cnidocysts, but the zoom is not capitate. The endoderm is leg are not represente(1 



scalariform and compact and thus the dactylozooid has no central lumen. The in this section. rf = dac- 



tylozooid, ^f = free por- 

 endoderm is on all sides surrounded by the supporting lamella and is not con- tion of the body _ wall of 



nected with the endoderm of the stolons. The nutrition of the dactylozooid from the gasterozooid, gs = 



gasterostyle. 

 the stolons must therefore take place through its ectodermal wall. — 



In spite of the very large material available for investigation I did not succeed in finding 

 the male colonies and determine their number of gonophores in the ampullae. Nor were young stages 

 of the female gonophores found in the numerous sections. The fully developed female gonophores 

 (PI. V figs. 46 and 49) have a complicated spadix. From a narrow base, where the main stolon enters 

 into the ampullae, the spadix spreads out semispherically; it is formed by numerous, fine, endodermal 

 blind sacs, the narrow and irregular lumens of which radiate out like the rays of a star from the centre of 

 the base of the gonophore. As the single blind sacs as a rule have a twisted course and not seldom 

 branch, the whole picture thus seems very complicated at first sight. The ripe ovum rests on the spadix 

 like one hemisphere on another and the spherical gonophore, which occupies the ampulla, is surrounded 

 by a thin endodermal layer. The colonies are not hermaphrodite and it is a puzzle even how fertiliza- 

 tion takes place in these animals with their closed ampullae. — As the ovum gradually develops to a 

 planula larva and grows in size, the spadix atrophies, its large cell material being probably used for 

 the nourishment of the ovum. When the spadix is more reduced (PI. V fig. 50) its structure be- 

 comes more distinct. — Other stolons of the wall of the ampulla may also secondarily come into 

 connection with the gonophore and thus increase the amount of nourishment during the embryonal 

 development. 



It has long been an open question whether the larva ruptures the roof of the ampulla when 

 it escapes. A series of sections through a number of large, empty ampullae points in an opposite 

 direction; in spite of the fact, that the larva must just have escaped from the ampullae, their roof is 



