28 



able to account more fully for the structure of these parts and the attachment of the mandibles. 

 When looking at the rostrum from its distal end (fig. 6 b), we see them through the inter- 

 mediate substance, like narrow lists in appearance, the free distal points of which are visible 

 in the mouth-aperture and are somewhat different in shape on the right and the left 

 mandible. Departing from the points, they turn outward, at the same time running down 

 the rostrum, their basal end lying inside its walls rather far from the aperture of the mouth. 



It is only the distal part of the rostrum which can be considered as formed by the 

 hypopharynx and the labrum, the proximal part must be chitine belonging to the ventral side 

 of the head itself, which here has become cone-shaped or forms the foot and the proximal 

 part of the cylinder. I draw this conclusion from the fact that the antennae, where they 

 are found, proceed from the basal part of the cone or from the foot {Choniostoma, pi. XI, 

 fig. 2 d), and that the maxillulae are situated on its lateral surfaces (see e. g. pi. X. fig. 6 a, c, 

 and many illustrations of heads of females seen from below). But these last-mentioned 

 mouth-organs must be treated separately. 



The MaxiUulce are found in all species. Each maxillula consists of a somewhat oblong 

 plate which almost throughout its whole length is coalescent with the middle and the more 

 proximal part of the rostrum, and in the latter place this coalescence is so complete that it 

 becomes impossible to distinguish the outline of the proximal part of the maxillula (fig. 6 a), 

 whereas its distal part (c) detaches itself from the lateral surface of the cylinder. Here it 

 divides itself into two branches, the anterior of which forms, now a shorter or fairly long, 

 now, and mostly, a very long process, which looks somewhat like a proximally very thick 

 and distally more slender seta. The posterior branch has a quite similar structure. These 

 two more or less setiform processes I consider as the principal branches of the maxillula; 

 they are never wanting, and as a rule they are somewhat curved (in the specially examined 

 specimen of Choniostoma their terminal half was sinuous), and on examining the head from 

 below, the anterior branch of the maxillula is mostly seen to proceed beyond the foremost 

 part of the lateral margin of the mouth-border, the posterior branch behind the posterior 

 part of the same lateral margin, whereas the distal part of its plate and the base of the 

 two branches are covered by the lateral part of the mouth-border, through which they can 

 be seen (e. g. pi. V, fig. 2d and especially fig. 3d). Besides, in most species the maxilla 

 possesses as an additional branch a process shaped like a stout and usually long seta, 

 articulated to that part of the maxilla which is coalescent with the rostrum, and often so 

 proximally that, in looking at the head from below, we get the impression that it is situated 

 outside the basis of the maxilla. The basal part of this additional branch is frequently set 

 off by an articulation. This branch is wanting only in Romoeoscelis and in the three species 

 of Sphceronella which are parasites on Cumacea, and which have no separate head. 



The whole rostrum is movable, so that its distal part with the moutli is turned 

 more or less forward or backward, now protruding, now receding considerably, which 

 differences are seen most distinctly by observing the head sideways, and comparing the 



