34 
THE VOYAGE OF H.M.S. CHALLENGER. 
B. The columellar muscle of Cavolinia longirostris (PL II. fig. 3). This 
muscle is not inserted in the middle line of the dorsal side of the shell but 
at the right hand angle of its truncature. 
C. The central nervous system of Limacina (PI. I. figs. 7, 8). We have 
already remarked that in this genus the ganglionic elements of the visceral 
commissure are united into two asymmetrical masses : the right large and the 
left small. This asymmetry is easily explained by the fact that the “ abdominal ” 
ganglion (the fused posterior visceral ganglia) has united with the “ supra- 
intestinal ” ganglion (right anterior visceral) ; the cause of the union is that in 
the asymmetrical Thecosomata it is the right-hand half of the visceral portion 
which is the best developed, that side of each set of organs having alone 
persisted. On examining the central nervous system of the Cavoliniidse it is 
seen on the contrary that the left-hand portion of the single ganglionic 
visceral mass, or the left half, if the two halves are distinguishable, is the 
larger ( Cuvierina , PI. III. fig. 1) ; this latter mass encloses the abdominal 
ganglion and the subintestinal (left anterior visceral), and in fact gives origin 
to the nerves of the “ abdominal ” ganglion, which arise in Limacina from the 
large right ganglion. Thus the abdominal ganglion has followed the viscera 
which it innervates (genital gland, kidney, heart, &c.) in their rotation from 
right to left (regarding the animal from the ventral surface). 
It is impossible, then, to deny the existence of this rotation converting the Lima- 
cinoid type into straight Thecosomata, or the descent of these latter from the coiled 
Thecosomata. 
In the Cymbuliidse the rotation has not been so complete as in the Cavoliniidse. The 
pallial cavity is not so decidedly ventral, its aperture being less symmetrical, a little 
oblique and more open towards the right side (PI. III. fig. 7), showing clearly that it has 
originated on this side and been displaced towards the left. Besides, the anus has been 
transported not so far to the left as in the Cavoliniidse, and is situated only a little to the 
right of the middle line. The shield (pallial gland), like the orifice of the pallial cavity, 
is still asymmetrical (PI. III. fig. 8), as in the Limacinidse, and not symmetrical, as in 
the Cavoliniidse. 
Supposing, then, for a moment, that the cephalic portion has remained immovable, the 
visceral portion of the Cymbuliidse has made a little less than a half rotation about 
its longitudinal axis. From this point of view, then, the Cymbuliidse are a little less 
specialised than the Cavoliniidse. 
The phylogenetic relations of the different genera of Thecosomata may be expressed 
graphically by means of the following table : — 
