308 DR. H. LYSTER JAMESON ON 
continuous with the bases of the epidermal cells, are attached to 
the ends of the muscle-fibres (mzse.), which may be produced out 
into tails. Tullberg recognised that such a junction was present. 
In Margaritifera vulgaris the attachment epithelium-cells may 
have one or several tails, probably also of connective-tissue 
character, passing over into as many muscle-fibres. 
Distally these epidermal cells broaden out somewhat, ending 
in a clean-cut surface, which may be represented in section by 
a clear zone forming a slightly acute angle with the sides. The 
nuclei, which are oval and about 3-4 uw long, are situated in 
the middle of their length. 
It is possible to make out, in some cases, a striation of these 
cells in the direction of their long axes. Occasionally the distal 
surface is raised into processes and papilla, but this may well 
be a result of imperfect fixation of the tissues. 
These cells stand out as stiff, independent, almost  bristle-like 
entities, and are probably hard and tendinous in character, 
They are frequently preserved in old preparations in which 
all traces of the structure of the ordinary epithelia have 
disappeared. In some cases they seem to have been drawn out 
in the fixing process ; thus the longest cell shown in Pl. XX XIIT. 
fig. 4a measured 26 4. It seems possible that in the shrinkage 
consequent upon fixation the majority of the elements here 
had broken away from the shell, but that this particular cell 
had remained attached and was consequently fixed in a state 
of extension. ‘This figure shows that the connective-tissue 
elements extend up between the bases of these cells. 
This epidermis is very closely adherent to the specialised shell- 
layer (hypostracum) to which it is attached, and the connection 
seems to be between the cells and the organic basis of the shell. 
Thus, in decaleifying a piece of the shell of J/ytilus with the 
adductor muscle attached, the hypostracum remained adherent 
to the epithelium and tore away from the rest of the shell 
CP]; SOXER Ve hie 75)) 
(11) SHett-SEcrerIoN. 
I will not attempt to survey the writings of previous inves- 
tigators on this subject. This has been ably done by Stempell 
(45), whose review contains a full and lucid discussion of the 
question. 
The general trend of opinion now seems to favour the theory 
dating back to Reaumer, 1709 (31), and held by Tullberg, 
Ehrenbaum, Moynier de Villepoix, and the majority of recent 
French and German investigators, that the shell (except the 
outermost layer of the periostracum and the hypostracum) is 
formed from a fluid secretion, rather than the theory specially 
associated with Huxley’s name (24) that the shell is derived 
from a succession of fully developed skins or cuticles, shed as 
membranes by the underlying epidermis. 
[50] 
