22 



knots may sometimes be present and sometimes absent, even in closely related 

 species; thus while they are found in Semiescharipora ovalis^ they are wanting 

 in Reptescharipora convexa". Secondly, we must remember that while the whole 

 surface of the colony is covered by a continuous membrane in the Eiithijris species, 

 every single zocecium in the mentioned fossil species has been furnished with 

 a separate frontal membrane, which has been situated below the roof formed by 

 the spines. As shown before, the marginal spines always spring from a Gymnocyst 

 and we must therefore assume that the Gymnocysl, which has formed the spines 

 of the marginal zocecia, has passed directly over into tlie basal wall of Ihe colony 

 from the free lateral edges of these zooccia. 



Rosette-Plates and Pores. 



The adjoining zooecia in a colony are, as known, connected by chords of 

 mesenchymatous tissue, which extend from one zocecium into another through pores 

 in the partition-walk, but while these pores appear in the Cyclostomatu as simple 

 perforations of the wall, they are in the Ctenostomata and Cheilostomata very fine 

 piercings in watch-glass-shaped, concave, thinned portions of these walls. These 

 peculiar formations have been noticed for the first time by Smitt, on the lateral 

 walls in Flustra foliacea^ and are called bj' him commimication-pores. He has 

 • not however seen the real pores, and seems to regard the whole, multiporous 

 plate as an opening. Later Reichert'' in a member of the division Ctenostomata, 

 Zoobolhrion pellucidiim, has observed the same formations, which he calls »Ro- 

 settenplatlcns and in which he has seen the real pores. Among later writers 

 Waters'' first drew attention to the importance of these rosette-plates for 

 the diagnosis of species and in a series of papers he gives information on their 

 presence in several Cheilostomata, while the writer of this work has illustrated 

 their occurrence in the Danish species, .^s this description'' was however written 

 in Danish, and for that reason less available, I may give here the results of 

 those older investigations lo which I have been able to add by later studies. 

 Though I do not find Ihe name rosette-plate good, I shall yet use it, jjartly in 

 view of its priority and partly because Waters has used it in his many papers. 

 The formations dealt with here may appear under two different forms, namely 

 as common rosette-plates or as pore-chambeis, and each of these can again be 

 divided into single-pored, or single and multi[)orous or compound. 



Rosette-Plates. A single-pored rosette-plate is a watch-glass-shaped, concave, 

 thin portion of the wall, which as a rule is surrounded by a more or less devel- 



' Sfi, Fl. 719; » S6. PI. 72U; ' 99, p. 426, PI. X.\. tig. 15; * 94. p. 2l>7 ; '■ 109, p. 286; •■■ 54, 55. 



