BRITISH NEMERTEANS, AND SOME NEW BRITISH ANNELIDS. 329 



A very interesting condition was found in two specimens of Tetrastemma vari- 

 color, which directly bears on the physiology of this region. In each a fragment 

 of the granular sac, with the central stylet attached, lay towards the anterior end 

 of the first region of the proboscis ; and since injury would scarcely have caused 

 a result so systematic, it is evident the stylet had been thrown off by the animal. 

 In both instances the central stylet-apparatus was complete, only in one the 

 anterior part of the basal sac appeared paler, and there was a slight irregularity in 

 its outline, similar to that in fig. 14, Plate V. In each, the lateral stylet-sacs had 

 their full complement of stylets, one or two of which equalled the central stylet in 

 size. There appears to be only one explanation of this state of matters, viz., the 

 fact that the central stylet can be thrown off, and somewhat rapidly regenerated; 

 for it is unlikely that in each case it found its way there from without, and it is still 

 less likely to have been driven in by an enemy. Former experience in regard to the 

 stylets from the lateral sacs shows that such bodies take some time to gain the 

 exterior of the worm, and hence our surprise is lessened at the perfection of the 

 new structures while the old have not yet escaped from the proboscis. Besides, 

 the structure of the parts in 0. pulchra will by-and-by throw still farther light 

 on this subject. 



Lastly, across this region passes the belt of granular glands (tt, various 

 figures), which have the form of lobules, with their long axes parallel to that of 

 the proboscis, and are situated beneath the two external layers of the part. The 

 granules are proportionally larger in Tetrastemma. I have not found any struc- 

 tural guide to their function, though they are invariably present in the Omma- 

 topleans. A curious appearance was noticed in a small specimen of Tetrastemma 

 varicolor, which had its stylet-region in front of the granular glands covered by 

 an external coating of large cells, with a nucleus and faintly granular contents ; 

 such, however, may have been due to an abnormality. 



The structure of the next division — that of the great Reservoir — has now to 

 be examined (g, Plate IX. fig. 3). On reaching the point previously mentioned 

 (a, Plate IX. fig. 11), the elastic coat and the external longitudinal muscular 

 fibres of the proboscis for the most part cease. Before this occurs, however, the 

 muscular fibres (t) peculiar to the region arise, sweep backwards in a beautiful 

 fan-like manner over the reservoir, loop round and meet those from the opposite 

 side, and leave only a small space in the centre posteriorly, through which the 

 channel of communication with the third region passes. When viewed as a trans- 

 parent object under pressure, or in longitudinal section, the direction of these 

 fibres is backwards and inwards. This great muscular mass does not receive 

 accessions from the outer wall, but the whole of the loops come from the front. 

 By the varied crossings of these fibres, a felted aspect is produced under examina- 

 tion in some species, such as 0. purpurea and 0. alba (Plate IX. fig. 11), and is 

 doubtless present in all. In addition, there are circular and longitudinal fibres 

 vol. xxv. PART II. 4 p 



