242 NATURAL SCIENCE. October. 



in the arrangement of the meshes may be traced, in the fact that on 

 the virgula-side (Fig. 13) stouter strands bend downward and run 

 round. These join the virgula at an acute angle. In the specimen 

 here represented they are only five in number. The proximal one is 

 somewhat coarser, and probably formed the mouth of the individual. 

 In the distal region of the initial canal, on the other hand, the meshes 

 are quite irregular. I have noticed the initial canal in only twelve 

 specimens, although I have had several hundred at my disposal. One 

 of these is complete, and has a length of 26 mm. A glance at Figs. 

 12 and 13 shows that the closeness of the meshes varies greatl^^ I 

 do not believe that the different sizes of mesh correspond to different 

 ages in the full-grown rhabdosome,^ but consider them simply as 

 instances of variation. 



With the ceasing of the initial canal the rhabdosome begins to 

 assume another shape. At the height of the second theca, from out 

 the confusion of meshes there appear at the angles four main strands, 

 which run through almost the whole rhabdosome. These main strands 

 divide the rhabdosome into four fields, almost at right angles to 

 one another : the virgula field, the two halves of which form an 

 obtuse reentrant angle ; the opposite plane or anti-virgula field ; and 

 the two somewhat convex thecal fields situated on either side. At 

 the same height as these main strands there arise in the middle of the 

 thecal fields, by each theca-mouth, discontinuous and somewhat 

 thinner longitudinal strands (Fig. 14). Each theca-mouth occupies 

 the whole width of the thecal field, and is therefore bounded at the 

 sides by the vertical main strands. Above and below the theca- 

 mouths are bounded by two horizontal cross-strands connecting the 

 main strands. The more meshes there are that occupy the intervals 

 between these few strands, the broader become the lower cross-strands 

 of the theca-mouths, till finally where the meshes are most compact 

 they grow into crescentic plates covering more than the whole theca- 

 mouth. 



The first theca does not lie in the thecal field, which in this 

 part of the rhabdosome hardly yet exists, but obliquely, so that its lid 

 (Fig. 13) is attached by one corner to the virgula itself. All specimens 

 with an initial canal, that I have seen, belong to fine-meshed in- 

 dividuals. 



As the meshes become closer they arrange themselves somewhat 

 differently. In the anti-virgula field they have no definite arrange- 

 ment, neither can any main strands of the second order be 

 distinguished. In the virgula field, on the contrary, they tend to 

 arrange themselves in four longitudinal series. This, however, is not 

 generally carried to any great extent, and one can hardly speak of any 

 new main strands. In the thecal fields this arrangement is more 

 evident. The strands are not hollow ; not even the virgula is that. 



1 Owing to an error this word was printed " rhabdome" in the first part of this 

 article. 



