OLIGOTRKMA PSAMMITKS. 265 



wall. As the branchial siphon is an invagination of the test, 

 the ectoderm is of course invaginated along with itj and as 

 the siphon has shrunk aAvay from the body-wall^ the ecto- 

 dermal epithelium remains adherent to the latter and forms 

 a lining to it in the space just mentioned. This same ecto- 

 dermic epithelium^ characterised by the presence of black 

 pigment granules, forms a complete lining to the cavity 

 marked x and to its posterior diverticulum. It may further 

 be seen that the cavity and the diverticulum are bounded by 

 numerous strong, longitudinal muscle-bands, which, as seen in 

 cross-section, appear to project into the cavities. The cavity 

 with its diverticulum, shown in figs. 7 and 8, is on the left 

 side; there is a similar cavity on the right side, with a much 

 shorter diverticulum, but as the sections are oblique it does 

 not appear in the figures. The ring of apparently thickened 

 tissue lying below the branchial sphincter shown in fig. 4 is the 

 external indication of the right cavity, and its walls are clearly 

 corrugated by the contracted muscle-bands which traverse 

 them. I am in doubt as to the nature of these cavities. They 

 may, of course, be artefacts formed by deep infoldings of the 

 ectoderm of the branchial siphon consequent upon the post- 

 mortem contraction of the muscle-bands. But I am not 

 inclined to interpret them as such, because of their sym- 

 metrical position right and left of the lower end of the 

 branchial siphon and the upper part of the branchial sac. 

 Nor can I imagine that any artificial folds due to post-mortem 

 contractions could give rise to so well defined a diverticulum 

 reaching so far back alongside of the branchial sac as that 

 shown in fig. 8. Furthermore the connective tissue of the 

 plerome adjacent to the cavities is differentiated, being far 

 more compact and containing more numerous cellular elements 

 than elsewhere, so as to give the appearance of a supporting 

 structure related to the insertions of the longitudinal muscles. 

 As it is impossible for me to form a satisfactory conclusion as 

 to the nature of these cavities from the examination of a 

 single series of sections, I must leave it an open question 

 whether they are artefacts or not; but I am disposed to think 



