EHABDITE-CELLS IK CEPHALOBISCUS. 261 



of any sort. They are not excretory organs, as their position and 

 structural relations imply, and they do not store up food-material 

 like the sacculi o£ the Crinoids, as the products of the " cells " 

 are (at least often) ejected on to the exterior. But since, on 

 the other hand, the ultimate products of these bodies are a 

 number of pointed or somewhat blunt rods, since the metabolism 

 of the " cells " is always in the direction of the production of 

 these rods, and since finally these rods can in some cases be seen 

 in the various stages of being, by the rupture of the " cell," 

 shot on to the exterior, the only structures with which the 

 " cells " can be compared are clearly the rhabdite- cells of 

 Turbellaria audTrematoda, and the less specialized bodies found 

 in the integument of Nemerteans. They will have the same 

 function doubtless as the rhabdite-cells in tlie latter groups, 

 whatever that is, and they are produced iu CepJialodiscus by the 

 following series of changes. 



It is first necessary to describe the structure of the wall of 

 the battery itself (fig. 1). Such a description is necessarily 

 based on the material as I found it, and is thus liable to a 

 percentage of error due either to post-mortem changes or 

 imperfect fixation or both. Figure 1 is a drawing of a portion 

 of the wall of the battery which, after a prolonged search, seemed 

 to have been most favourably preserved and cut accurately at 

 right angles to the surface. The cavity of the battery (10) is 

 lined by a series of occasionally nucleated fibres (<§), which are 

 doubtless the direct continuation of the longitudinal fibres of 

 the stem described by Mcintosh. Many of these fibres have 

 snapped in fixation (9), giving the appearance under ordinary 

 lenses of a row of large cilia projecting into the cavity. Situated 

 on these fibres is a single row of large irregular cells {4-), each 

 containing at least one undoubted nucleus. In some preparations 

 it can with certainty be made out, as shown in the figure, that 

 the cells are continuous at the base, so that a perfectly con- 

 tinuous layer of protoplasm surrounds the longitudinal fibres (5). 

 From this layer of protoplasm there occasionally passes a long 

 filament (7) which lodges one or more nuclei, and passes straight 

 upwards to anastomose with the free surface of the battery (-?). 

 Similar filaments, which are however much more numerous, and 

 also lodge nuclei, pass from the basal cells themselves to the 

 surface {6). Whether either one or both series of filaments 

 represent the narrow interstitial cells described by Biirger as being 



