414 G. HERBERT FOWLER. 
form shells, including the Phascolion strombi (Sipunculus 
bernhardus, Forbes) of our own coasts. In the case of several 
species, it has been noticed that they tenant either a heliciform 
or a straight (Dentalium) shell. 
To summarise—we know (1) that three genera of similar 
and characteristic form (though not closely allied) contain a 
spiral chamber inhabited by a Gephyrean; (2) that in the 
case of some species of all three genera, this chamber is 
certainly due to settlement on a heliciform shell; (3) that in 
the latter case the characteristic shape is nevertheless acquired 
by growth over the shell; (4) that many species of Gephyrea 
normally inhabit heliciform shells irrespective of the presence 
of Madreporaria; (5) but sometimes tenant a straight one, so 
that curvature is not a primary physiological necessity to them. 
To these may be added (6) that in all the species, if settlement 
is known to occur at all, it is always on a Gastropedan, never 
on dead coral or rock. I would therefore submit that, in the 
absence of experimental proof, we are for the present justified 
in regarding the heliciform cavity as due, in those cases where 
we have no direct evidence, to the same cause as in those 
where our knowledge is more exact. 
Professor Semper’s suggestion, that a faster growth of the 
Gephyrean in length than that of the coral in breadth pro- 
duces the spiral chamber, seems inadequate, unless the worm 
possesses the power of boring into or absorbing the hard 
coral, a power not, I believe, attributed to it. 
I have been permitted to grind a microscopical section of 
the corallum of Heteropsammia Michelini in the British 
Museum; and though no trace of a shell is visible round the 
spire, yet the many admitted cases in which various organisms 
absorb a Molluscan shell (for example, a Hydractinia, some 
Porifera (and Adamsia?), from our own coasts can entirely 
absorb the Buccinum shell on which they settle) deprive this 
argument of all force. 
The anatomy of the colony in the main resembles that of the 
Rhodopsammia previously described (“ Anat. Madr.,”i). Colony 
and polyps are alike clothed in a continuous body wall of the 
