Br. F. A. Bather — Notes on Yunnan Cystidea. Ill 



An outer surface that has been slightly worn, so as to uncover the 

 cy-chanrrels without obscuring them, shows the outer openings of the 

 pore-canals connected by shallow channels on the following plan 

 (fig. 15). The simplest arrangement consists of two channels, curved 

 like the sides of an 0, with a pore at each end. Next comes a 

 doubling of these channels. Both of these arrangements are well 

 shown near the base of E 7676 (fig. 16). It appears as though the 

 doubling were due to, first, a broadening of each channel, and, 

 secondly, a median up-growth of its floor. This process may take 

 place on one side only, so as to yield a total of three channels; or it 

 may be repeated on one side, yielding five channels (fig. 15), or, very 

 rarely, on both sides, yielding six. 



It is, however, unusual to find quite so simple a pattern : modifica- 

 tions arise in two ways. First, by simple irregularity in the curves 

 of one or more channels. Secondly, by other up-growths of the floor 

 so as to bar one or more channels (fig. 17). Each of these modifica- 

 tions seems to be subject to a special limitation. The irregularity is 

 subject to the bounds which are set to the channel-system of a single 

 pore-pair. The barring is subject, apparently, to the condition that 

 such a channel-system must be continuous ; thus, no part of a 

 channel has been observed with a bar at each end ; however irregular 

 the labyrinth may become, it is always possible to track along every 

 part of it. 



Now as to the bounds of a channel-system. In no case does 

 a system transgress the boundary of a plate ; no pore-canal, and no 

 channel, ever crosses a suture. Within each plate the diplopores are 

 distributed over the whole surface — not regularly or according to any 

 pattern, but at approximately equal distances. There are two plans 

 of structure: the diffuse (figs. 16, 17, 18) and the concentrated 

 (figs. 19, 20, 21). In the former the channel-system is diffuse 

 and is separated from adjoining systems by a tract no wider 

 than a single channel or, wbat comes to the same thing, no wider 

 than a ridge between two channels. No wider is the space left at 

 the sutural margin. Each channel-system spreads itself out until 

 brought up against its neighbours; and, since the foci of each system 

 are distributed irregularly, the boundaries also become irregular; 

 did the systems start from equally - distributed centres, their 

 boundaries would form regular hexagons, but, as it is, they form 

 irregular polygons. It is therefore not easy to distinguish these 

 boundaries from the other ridges, and so the whole thecal surface 

 appears at first sight covered with a confused maze of channels. 

 It seems possible that this complexity has misled even so acute an 

 observer as Professor Jaekel (1899). His plate iv, fig. 2a shows in 

 some cases as many as 5, 6, or 7 pores belonging to the same channel- 

 system. There may in rare cases be more than one pore-pair to 

 a system, but I am not convinced even of that. What does often 

 happen is that the outer opening of one (or both) of the pores is 

 crossed by a bar (fig. 17), so that externally the pore seems duplicated 

 or even triplicated. There is also a deceptive appearance of pores at 

 other bars. 



Finally, the pores and their connecting channels are covered on the 



