Behavior of Tubicolous Annelids 179 
tors in matters of behavior. ‘They are not of sufficient importance 
to warrant any explicit deductions, but are simply given as facts 
more or less directly related to phases of metabolism and behavior 
presumably more or less intimately correlated. While the effects 
of the supercharged water with CO, were not unexpected, those 
associated with the reduced oxygen content were something of a 
surprise. Both series of experiments would seem to suggest that 
certain organisms are capable of enduring inimical respiratory 
conditions which would be quickly fatal to others. “They further 
suggest that the signs of distress sometimes observed in these 
creatures in the aquarium in which the water has become “bad” 
may not be due to lack of oxygen, not to the presence of excess 
of CO, but to toxic organic matters with which the water has be- 
come charged. 
MODIFIABILITY OF BEHAVIOR 
While attention was directed to this feature in the earlier et 
(pp. 304, 316), no special details were given. Early in July I 
obtained from Mr. Geo. Gray a colony of F ydroides which had 
been kept in his aquarium over the preceding winter. It seemed 
probable that some evidence might be obtained concerning the 
effects of long confinement under artificial conditions which might 
be of importance. The aquarium in which they had been kept 
was in the basement of the laboratory, and was but dimly lighted. 
It was, however, near one of the working tables about which 
there was more or less passing. A careful examination of the 
colony seemed to show that it was apparently normal, and in an 
_ average state of vigor so far as could be observed. On testing 
by the shadow stimulus it was found, however, that they were 
strangely negative, only a specimen here and there showing any 
response to the tests. [he colony was taken to my private lab- 
oratory where they were placed in the most favorable conditions 
for accurate observation. Subjected here to the usual tests, at 
first not more than about 10 per cent responded in the usual man- 
ner. And after the first test or two of a series even this ratio de- 
creased almost to the vanishing point. [wo or three possible 
explanations may besuggested. First, that we havehere a counter- 
