Behavior of Tubicolous Annelids 183 
ina, Illyanassa, etc., and colonies occur also on the chelz of the 
lobster, and doubtless many other such habitats. It can hardly 
be seriously contended that in such living habitats where they are 
constantly subject to change of position and relation any such 
factor as gravity or light has been dominant. 
But one of the most interesting of habitats is that of the nets 
forming a part ofa fishtrap. It is of course well known that these 
traps are set up in water of from ten to twenty feet in depth, and 
are continued in use for some three or four months during the 
summer. In this particular case the tubes were taken from about 
the middle of the net, which would again preclude the direct 
action of either light or gravity in the matter of attachment. We 
have the further point of interest that some idea of the rate of 
growth may be inferred, since 1t must have taken place within the 
few months of a given season. 
In this connection occasion must be taken to refer to a few fea- 
tures of behavior shown by Potamilla and Sabella. Since the 
observations noted in the former paper I have taken occasion to 
repeat and extend them. An examination of many specimens 
as they occur in nature failed to show any definite orientation to 
either light or gravity. As is well known these species form a 
tough, flexible tube, which may be easily bent in any direction. 
Loeb has found that Spirographis, an annelid having a similar 
tube, but growing to large size, shows a definite response to light. 
I have kept species of both Potamilla and Sabella in the aquarium 
for weeks, variously disposed as to light, but was not able to dis- 
tinguish any definite evidence of such orientation. 
CONCLUSIONS AND REFLECTIONS 
On the basis of the so-called tropism theory, using this term in 
its current sense, these rather remarkable variations and individ- 
ual differences of behavior are difficult to explain. If conditions 
of density, temperature, oxygen, etc., are to be regarded as pri- 
mary determining factors, how with these more or less constant 
should behaviour differ; or with these differing to marked degree 
why should behavior continue more or less constant? Again, 
