2f) 



roselle-jilales. While Xilsche' in his nl)ove-inenli()iu'(i work on Meinhr. mein- 

 branncca has rightly ohserveii, that the single zorecia have independent lateral 

 walls, he gives a wrong view of the relation of the single zoa^cia to the rosette- 

 plates. He saj's namely: »Die Rosettenplatten eines jeden Zoociuni eorrespon- 

 diren nun mit den Rosettenplatten der uniliegenden Zoocien auf das genaueste*, 

 and he gives in detail an explicit account of how the rosette-plales of each 

 zoceciuni are placed opposite to a corresj)onding rosette-plate in an adjoining 

 zooecium. If for instance we separate a row of zod-cia of a ScriijxH-clldrid form 

 (PI. II. figs. 7 g, <S c) each lateral wall in its distal half will show a inultijjorous 

 rosette-plate, hut in its proximal half an opening of the same sliajje and 

 size. If we subject Flustra foliacea (PI. I, fig. 81)) to the same treatment, we find 

 2( — 3) multiporous rosette-plates on its distal half, and '2{- \i) openings on its 

 proximal half. Because of the arrangement of the zonecia in alternating longitudinal 

 rows, one or more openings in the proximal half of a lateral wall will always 

 correspond to and fit exactly o))posite the same number of rosette-plales in the 

 distal half of the corresj)onding lateral wall of the adjoining zooecium. In (ie- 

 mellariu loricata we have an example of a form, the zod'cia of which in contrast 

 to the ordinary conditions are arranged in pairs. Each two of these zoa-cia are 

 as a rule connected by a single rosette-plate, which only belongs to the one 

 zoa'cium, while the opposite one has a corresponding opening in the wall. We 

 may examine ever so many forms in this regard, but we will never find two 

 rosette-plates opposite each other, but a rosette-plate on one wall always corre- 

 sponds with an opening on the ojjposite wall. Strictly speaking the rosette-plate, 

 as well as the apparently single lateral wall between two adjoining zocecia, is 

 also divided into two halves (PI. \\'II, tig. 10 b), which however in the case of 

 the rosette-plate are very unequal in size, as the concave pore-bearing portion 

 belongs to the one wall, while the oj)|)osite wall includes Ihe pore-iing, which 

 can then be seen on the inner (towards the inside of the corresj)onding zoo-cium) 

 surface of this wall as a more or less circular projection round the above-men- 

 tioned opening. The rosette-plates are arched inwardly towards the proximal 

 zooecium on the terminal partition-wall, which as already said is common to 

 two zocecia situated behind each other. The above-discussed arrangement of the 

 rosette-plates can be illustrated in a very clear way by means of a variety of 

 Flustra seciirifrons" with narrow branches found in the Kara Sea, in which the 

 rather numerous uniporous rosette-plates are unusually strongly arched, so that 

 they can be seen llnough the membranous oral wall. Inwardly arched rosette- 



^0 ».i/i, 



53, I'l. .\.\V1, lig 9. 



