ON THE MICKOSCOPIC STRUCTURE OP SHELLS. 
]01 
layer. In Castalia this may always be distinguished where the epidermis is 
eroded near the umbo; a dead-wliite surl'ace being seen where the erosion 
^ not laid bare the nacre. In fact, as regards the structure of its shell, 
does not differ from the true Nayadac ; and by this character it 
should be united with them. VVe have here another interesting example of 
agreenient between the intimate structure of the shell and the general con¬ 
formation of the animal; since Deshayes and other recent systematises have 
api^d the decided opinion that the genus Caslalin must be merged in 
Into, as no valid distinctive characters can be assigjjed to it. In Triffonia, 
on the other baud, the cellular layer is very thin, and the cells, instead of 
being polygonal, have their angles rounded ofl', very much afur the manner 
of those which lie inunediately beneath llie epidermis of Vmo (fig. 51, first 
Kc{wrt*). The internal or nacreous layer of Triijonia jiresents a few scat¬ 
tered and irregulw tubuli; but from their minuteness and want of definite 
arrangement, their existence can scarcely be regarded as characteristic of it. 
VIl. Arcacc<e, 
73. In this family I have examined shells from the genera Area, PcctuH' 
and iVwf«/a; the conformity in the structure of which is generally 
very close. The texture is usually the same, or nearly so, in tho iuner and 
be outer layers, although their colour may differ, as in Area No<c or Pec- 
^nctiluspilosus. The shell contaius but a small proportion of the mem- 
ranous element ; and no other definite arrangement of its coiupnnents can 
e made out, than is indicated by that kind of corrugated aspect already de- 
senbeo. In many species of Area and Pttlunmfus there is a very' regular 
■** tubuU commencing on the internal surface of the shell, 
in the uitervals between the ribs, and spreading out as they pass towards tho 
fK \ so as to meet each other from opposite sides of the ribs 
The diameter of these tubes varies from about l-7000tb tol-5000th 
?, imm. All the specie* of Area, in the shells of which 1 have met with 
eni, namely, A. Note, A. iaeftra, and A. barhuta, belong to that division of 
epnus in which the valves gape for tiie pa.<«sage of the byssus, or rather 
® I ® *™dii3ou» prolongation of the foot, by which fho animals attach them- 
The only species I have examined of that division in which the 
completely, is A. senilis, in whicli tlie tuhulai' structure is alto- 
6 er absent. Should this difference be found to exist constantly between 
respectively belonging to those two groups, it will he another 
general structure of tlie animal 
Nuab furnished by the minute organization of the siiell. In 
mwbM inner layer has the tnily nacreous stnicture, as 
luhiti fruni it* iridesceut lustre. In the outer, a small amount of 
thpir**' ***'^®*'***^^ ^ observed; and here, as elsewhere, tho tubuli in 
outwards seem to be directed towards the ribs, so as when they 
the rf to leave the spaces between these altogether free, whilst in 
tubular layer the tubuli are chiefly found in these 
In fig 15 shown a portion of the <!Xlcrnal layer, the section being 
the j ® ®**’'Taee as possible; the tubuli arc seen to open upon 
bands da, 45, which correspond with the ribs of the shell, whilst 
S' cell? have, been dweribed (Transactions of tlio Microscopical 
**■ r) “ l«l«nnng to Uie periostraciim itselt His whole acnjimt of them, 
fKhibh irk “"e»l>ond» with the appeanincrs which the pcrioBtracum of I mo will 
of tt». t?.“ i* the shell by the ageocy of acid, hnt carries with it a por- 
P®sition as tlie Mircrt' *' ^ hesitation in regarding my own description of their 
