Br. F. A. Bather — Notes on Yunnan Cystidea. 319 



the latter species. Similarly such specimens aB E 7631, E 7635, 

 E 7638, and E 7644, each with 5 plates, cannot otherwise be 

 distinguished from E 7636 with 6 plates; while E 7630 and E 7634 

 with 6 plates are otherwise the same as E 7633, E 7640, E 7642, 

 E7677, and E 16168, each with 5 plates. 



There are a few exceptions to the preceding general statements. 

 S. A. Miller (1879) regarded the additional plate in If. baculus as an 

 AdorallV rather than III, but, since his figure lends no particular 

 support to this view, the species has been included in the above liBt. 

 In M. gyrintis (fig. 30) it does seem probable that an Adoral IV 

 enters into the periproctal frame, but it is obscured. In 

 M. splicer oidalis (fig 27) there are at least two additional plates in the 

 Ad. Ill series; indeed the drawing shows three, but the small 

 triangular plate is possibly a corner separated by a crack. Thus the 

 periproct is surrounded by 7 or 8 plates; but this extra number is 

 correlated with the spheroidal shape of the theca rather than with 

 any structural peculiarity in the periproct. 



There may also be a reduction in the number of plates. The 

 specimen figured by It. It. Rowley (1903) as " Holocystites 

 papulosus?" shows the posterior Adoral II descending on each side 

 of the periproct so as to meet the two Adorals III and to exclude the 

 facetted plates from the periproct, which therefore is surrounded by 

 only three plates. By a similar downgrowth in M. subovata the 

 right posterior facetted plate is excluded from the periproctal frame, 

 which here consists of four plates. This is an exaggeration of 

 a tendency which this species shares with M. ornatissima (fig. 25) 

 and M. papulosa, and which results in an obliquely elliptical periproct 

 with its long axis passing from the left facetted plate to the right of 

 the two Adorals III. In the otherwise similar M. aspera thi& 

 tendency is not manifest, so that the long axis of the periproct is 

 horizontal. 



The numerical reduction discussed in the preceding paragraph is 

 due to, elimination of a plate or plates from the periproctal frame. 

 Specimen E 7632 presents a reduction to four plates owing to the 

 presence of only one Ad. Ill in the frame; and this one, being 

 a large plate, may be taken to represent the usual two, whether 

 there has been fusion or no. In other respects this specimen 

 resembles E 7631 and E 7635, and the reduction can only be 

 regarded as a meristic individual variation. 



Other exceptions to the general statement are caused by modifica- 

 tions in the posterior interradius of circlet Ad. II. Thus, in E 7639 

 the left facetted plate is excluded from the periproctal frame by the 

 intercalation of a plate between it and the posterior Ad. II, so that 

 there are 9 Ad. II instead of 8 ; correlated with this is an extra 

 plate in circlet Ad. I, apparently cut off from the left posterior Ad. I. 

 This intercalation of plates is, oddly enough, not accompanied by an 

 additional Ad. Ill, so that the number of plates surrounding the 

 periproct is still 5. In E 7675, a somewhat similar intercalation 

 of an Ad. I and an Ad. II only just eliminates the left posterior 

 facetted plate from the periproctal frame, and is accompanied 

 by an additional Ad. Ill, so that there are 6 plates in the frame 



