320 Dr. F. A. Bather — Notes on Yunnan Cystidea. 



and only just escape being 7; in this specimen, however, the 

 three posterior Ad. Ill are relatively much narrower than in 

 the more swollen or spheroidal E 7639. In these cases the additional 

 Ad. II seems to be due to nothing more than the development of 

 a suture separating the prolongation of the left posterior facetted 

 plate from the body of the plate, for the sake of greater flexibility ; 

 in harmony with this the adjacent Ad. I also becomes longitudinally 

 divided, but its suture in both specimens is a very close one and not 

 easy to see. 



In no specimen of Megacystis recorded or known to me are the 

 anal valves (or periproctals) preserved. The usually rectilinear 

 outline of the opening, with its slightly rebated sides, denotes that 

 the periproctals were, as in Sinocystis, triangular plates. Their number 

 may be inferred from the number of sides possessed by the periproct. 

 This is not necessarily or invariably the same as the number of 

 bounding plates. Occasionally the sides of the periproct, and 

 consequently the periproctals, do approximately correspond with the 

 margins of the bounding plates in both position and number, e.g. 

 E 7634, and Miller's figures of If. parva and If. scitulus, his fig. 6 of 

 If. commoda (not the figure copied antea fig. 23), and less closely his 

 figure of II. gorbyi. Frequently the sides correspond with the 

 plates in number but not in position, and in this case the angle at 

 which two sides meet lies not at the suture between two bounding 

 plates but at some distance from it, 60 that the free margin of the 

 dlate is notched. This notching may affect one or more or even all 

 plates ; thus E 7640 and E 16168 have each 5 sides and 5 bounding 

 plates, and all the latter are notched. Usually, however, the 

 correspondence is fairly exact with Adorals II ;it is the Adorals III 

 that tend to be notched. When the numbers correspond the notching 

 maybe confined to the left-hand Ad. Ill (e.g. E 7642, E 7645,E 7677); 

 when there are 6 sides and 5 plates, the notching extends usnallv to 

 the right-hand Ad. Ill (e.g. E 7638, E 16167) or it may instead affect 

 the left Ad. II (as in E 7637). In the cases (probably rare) when 

 6 plates surround a 5-sided opening, the median Ad. Ill has a con- 

 spicuous median notch (e.g. E 7636). Sometimes it is impossible to 

 identify any precise number of sides, since all angles seem rounded 

 off so as to produce a circular or elliptical opening. This is 

 particularly noticeable in the If. ornatisswia series (see p. 257) and 

 in If. sphceroidalis. Possibly in such forms the triangular periproctals 

 were separated from the periproct margin by a flexible finely-plated 

 membrane. In U. aspera and in the specimen described by Rowley 

 as " Ifolocystites papulosus ? " (see p. 257) the periproct has a quad- 

 rangular outline and may have been closed by four valves. 



The preceding facts show that, though there were certain tendencies 

 in the several groups of species, still there was no fixity. Neither 

 the number of the periproctals nor that of the sides of the periproct 

 can be regarded as a specific character. Possibly the sides of the 

 periproct may originally have corresponded with the margins of 

 the bounding plates, and this correspondence is usually retained in 

 the adoral half of the periproct ; but the variability in the third adoral 

 ■circlet may have conflicted with the persistent conditions in the 



