595 
tion and plaster the dorsal shell layers against the dorsum of the 
metanepionic substage? This is partly answered by the fact that 
the tendency to shell building on that side would prevent this until 
a small umbilical perforation was formed and also by the fact that 
in many shells the whorl of the paranepionic actually does plaster 
itself on to the dorsum of the metanepionic and the umbilical per- 
foration is reduced to a very small aperture. It is, however, abso- 
lutely essential to call in the aid of heredity, otherwise the tendency 
to shell building in the dorsum of the nepionic stage cannot be 
considered sufficient to prevent the entire obliteration of the umbil- 
ical perforation. ‘The shell on the dorsum of the older stages is in 
great measure absent in most nautilian shells, but there is no such 
difference in gyroceran or cyrtoceran forms or in loose coiled gyro- 
ceran form with the whorls touching, nor yet in nautilian form with 
very faint contact furrows. ‘The tendency to build thick shell on 
the dorsum of the nepionic whorl, while still free, is therefore one 
that can only have been derived from shells having free dorsal sides 
and this tendency is obviously strong enough to stiffen that side 
and prevent the entire closing of the umbilical perforation. 
In the second place there are a number of Cretaceous, Tertiary 
and recent Nautiloids having accelerated development of the dorsal 
furrow, and in these the furrow appears on both sides of the com- 
ma-shaped umbilical perforations. It is perfectly plain in these 
that no bending of the whorl could account for the result and that 
it is in no sense due to a moulding of one whorl upon another. 
The outlines of the dorsum of the paranepionic substage in these 
species does not coincide with those opposed to it; they are the 
reverse of each other. 
Nevertheless it is practicable, as has been said above, to appeal 
to the curvature of the paranepionic volution at the gyroceran bend 
as a possible mechanical cause for the incoming of the dorsal furrow 
on the distal part of the curve so long as the curve is sufficiently 
abrupt to produce it, or so long as the absence of ancestral forms 
does not enable us to trace the origin of this character back to a 
contact furrow and account for its presence in the earlier stages of 
species like those of the genus Tarphyceras by the action of the law 
of acceleration. 
I have consequently thought it safer for the sake of argument to 
concede that the dorsal furrow of Tarphyceras was perhaps present 
