88 FORMATION OF SHELLS OF ANIMALS, ETC., 
the spaces existing between the longitudinal portions of 
dentine. These latter spaces are in principle analogous 
to the canaliculi of bone, which will be fully explained in 
the article on bone. Hence, in one form of bone, the 
cementum or crusta petrosa canaliculi and dentinal canals 
sometimes exist together. The dentinal canals are merely 
spaces of feeble or imperfect cohesion, continued from the 
interglobular spaces, where such exist, between longi- 
tudinal portions of coalesced dentine to the pulp-cavity. 
These passages, becoming a little widened from the con- 
traction of the part in drying, and containing air, seem 
to have, under the microscope, especially if appearances 
depending upon distance are not sufficiently allowed for, 
the appearance of tubes. It is, however, incompatible 
with the function of tubes, as they exist mm other parts, 
that a system of such organs, intended for the conveyance 
of secreted fluids, should be simple prolongations of the 
interstices of the body, as are the spaces between globular 
portions of dentine and the pulp-cavity of a tooth, this 
latter, notwithstanding its size, being only a cellular 
‘interval. The other structure is distinctly tubular. It does 
not exist, excepting in the tegumentary part of the shell, and 
it is this which is most unquestionably represented in the 
plate before referred to. These tubes pass from the external 
surface of the shell, through its substance, to its deep one. 
They have distinct parietes; and their peripheral extre- 
mities in different parts of the same shell present very 
different forms. All along the convex border of the claw 
each of these tubes projects beyond the surface, where it 
becomes free, presenting generally a feathery extremity, 
or sometimes it seems split up into small fibres like a 
brush. (See fig. 5 a, a.) Just at the point, and all 
along the biting edge, these tubes are rather smaller and 
ee 
oor 
