36 
THE VOYAGE OF H.M.S. CHALLENGER. 
been accompanied by the loss of the operculum ; in the large adult Limacinse (in Limacina 
helicina, for example) the tendency is already observed to lose the operculum, and even 
the partially uncoiled embryonic shell of Gleba has none, nor have the embryonic shells of 
the Cavoliniidae. 
3. The most primitive of the straight Thecosomata (subgenus Creseis of Clio ) are 
circular in section, and have retained certain traits of the Limacinse ; thus Clio virgula 
has the initial part of the shell coiled dorsally, indicating a former coiling. These forms 
of Clio have also on the dorsal margin of each tin the same small tentacular lobe as is 
found in certain Limacinse. It is easy to see how the forms compressed dorso-ventrally 
and with lateral keels ( Clio properly so called) have originated from these species of 
Clio which are circular in section. 
4. From these latter the Cuvierinse have arisen by constriction behind the aperture, 
and by the formation of a diaphragm about the middle of the shell, behind which the 
initial part of the shell may become lost. 
5. Finally, those forms of Clio which are compressed dorso-ventrally, in which (as in 
Clio cuspidata, for example) the lateral margins have diverged almost in opposite direc- 
tions, the aperture being produced into a narrow slit at either side, have given rise to 
the most primitive Cavolinise, which, as in the case of Cavolinia trispinosa and Cavolinia 
quadridentata, still retain some of the characters of Clio or other primitive characters, 
such as a well-marked embryonic shell, fins distinct from the rather narrow ventral lobe 
of the foot, ganglionic elements of the visceral commissure still separated into two halves, 
&c.; and from these forms the passage is easy to al] the other Cavolinise. Embryology 
confirms the view that the Cavolinise are the most specialised in this sense ; in fact 
Fol 1 remarks in this connection, “. . . les Hyaleacees sont un extreme.” 
This account of the phylogeny of the Thecosomata, drawn up from comparative 
anatomy and based upon embryology, is also found to agree with palaeontology, which 
gives it additional support. We find Limacina in the most ancient Tertiary deposits, 
and also forms resembling Clio, with circular transverse section (Euchilotheca) , as well 
as nearly related forms which lead on to Cuvierina ( Tibiellct ). Clio, properly so called, 
however, and Cavolinia, do not appear until the Miocene. As for the Cymbuliidae, it is 
hardly possible that their “ cartilaginous shells ” should be preserved. 
As regards the fossils considered to be Primary Thecosomata ( Creseis , Cleodora, &c.), 
and the larval shells of Cymbuliidae described by Ehrenberg, we shall soon see the slender 
basis on wdiich rests the systematic position assigned to them. 
From what has been stated above, it is easy to see the importance of the position 
occupied by the Limacinidae in the morphology of the Thecosomata. By means of the 
knowledge of Peraclis they enable us to understand the relations of the Cymbuliidae, 
which were extremely difficult. Wagner 2 even considered them as the most primitive 
1 Loc. cit., p. 206. 2 Die Wirbellosen des weissen Meeres, Bd. i. p. 119. 
