APPENDIX — SPONGES OF THE "AKGO" CKTJISE. 215 



sphere (or sphere?), similar to Bowerbank's* figure of the 

 rosette in Esperella lingua, B. ; secondly, sigmata, - l by 

 0"006 mm., generally simple, a few also contort ; thirdly, 

 toxa, varying greatly in size, from O'l to 0*25 mm. by 

 0-002 to 0-005 mm. 



A number of most interesting embryos were found in 

 this specimen. They are generally of an oval shape, 0"3 

 by 0'21 mm, the one pole (ventral) being slightly broader 

 and more flattened than the other one (the dorsal pole). 

 The outer wall, except on the ventral pole, consists of 

 high and exceedingly narrow columnar cells (the ectoderm) 

 with the nuclei quite at their bases. Since a section 

 through such an embryo is bound to be thicker than the 

 diameter of its ectodermal columnar cells, and since the 

 embryo is round, we must of course expect to se3 several 

 rows of ectodermal nuclei, which will be the more numer- 

 ous the thicker the section is. This is shown in the 

 section represented in PL XI, fig. 9. Cilia or fiagella 

 could not be detected on the ectoderm. The ectoderm is 

 absent on the ventral pole of the embryo. The central 

 mass of tissue which also forms the ventral pole of the 

 embryo, seems to be gelatinous, and contains numerous 

 nuclei which are larger, but less intensely stained than 

 the ectodermal nuclei. This tissue also contains spicules 

 and for this reason one is inclined to take it to be 

 mesoderm ; so what part of the embryo represents the 

 endoderm is a mystery to me. These embryos already 

 possess a skeleton, both megascleres and microscleres. 

 The megascleres are styli, 0*184 by 0'0028 mm. They 

 are arranged in a sheaf, converging with their pointed 

 ends and directed towards the ventral pole of the embryo, 

 and diverging with their blunt ends and directed towards 

 and beyond the centre of the mesoderm (see PL XL, fig. 8). 



*Bowerbank, British Spongiadte, vol. i., pi. xviii., fig. 297, 



