ORGANIC DEPENDENCE AND DISEASE 
53 
eom. There are long tubular passages between the coral- 
lites in early-growth stages which have not been described 
in the structure of this coral genus but undoubtedly belong 
to it. In sections these may be confounded with worm tubes, 
but in etched specimens, such as have here principally 
served for illustration, tlieir real nature seems to be clear. 
In this interesting combination there is still another 
member — a small silicious sponge. It has come to my 
notice several times. The one here figured was taken from 
the empty tube of the worm, but whether that is its usual 
position or whether it may seat itself in one of the coral 
calyces or whether indeed it is a usual member of the con- 
sociation cannot be regarded as clearly established. Its im- 
portance is not to be magnified; such little organisms are 
easily entangled in growing corals and must be expected in 
the fossil state. 
Some illustrations are here given which show how readily 
the dead parts of these organisms become encrusted with 
serpulid worms. Figure 29 is the surface of a part of a 
dead Loxonema to which a Pleurodictyum had grown, and 
figure 32 shows the inside of an old tube of the commensal 
worm (which is known as Hicetes innexus ), itself incrusted 
with minute worm tubes. 
Interesting as is this instance of commensalism, its most 
extraordinary feature is the evidence of selection by the 
larval coral, of the body which serves as the base on which 
it is to grow. It is stated above that a very evident ma- 
jority of the colonies of this coral Pleurodictyum, as it oc- 
curs in the Hamilton shales, are attached to an organic 
object and that this organic base in apparently a very 
large majority of the cases is a shell of Loxonema hamil- 
toniae. Occasionally the shell may be a Pleurotomaria of 
one or another species. On the other hand the Rhenish 
Pleurodictyum problematicum fixes itself by decided ma- 
jority to the brachiopod Chonetes sarcinulatus Schlotheim. 
I have examined a considerable number of specimens 
