210 
BULIvETIN of the bureau of fisheries. 
taining at least traces and sometimes significant quantities of water-gas tar and other 
oily matter. The results of this investigation indicate, in a word, that the tar or oily 
wastes could have no effect on shellfish in this way. The details of this part of the work, 
however, are reserved for a report dealing specifically with the effect of water-gas tar on 
oysters. This paper is confined to oxygen requirements of shellfish and their resistance 
to lack of oxygen. 
METHOD OF EXPERIMENT. 
The method of manipulation was, in brief, to place the shellfish in a desiccator 
completely filled with sea water of known content of dissolved oxygen, leave the apparatus 
at some constant temperature during a definite period, and then to sample the water 
in the desiccator so as to compute its decrease in dissolved oxygen. Winkler’s well- 
known titration method was used to measure the oxygen both at the beginning and end 
of each experiment. A vacuum desiccator was used for the containing vessel because 
it could easily be closed water-tight, could accommodate almost any size of shellfish, 
and enabled one to take through the side opening with glass stopcock a fair sample of 
the contents. 
In practice a number of precautions were found necessary. The hollow dome of 
the desiccator cover was entirely filled with paraffin to exclude air from entrapment 
in it. A glass tube within the desiccator was connected to its side stopcock and reached 
nearly to the bottom. When, therefore, the filled desiccator was opened at the top 
water could be sucked off through the opened side cock into the sampling bottle so that 
the sample would come from a point well below the surface of the water, where the oxy- 
gen content would be fairly constant. The sea water ® used in each experiment was 
brought to the required temperature and placed in a large reservoir jar on a shelf, from 
which it could be siphoned without bubbling into the desiccator. When it was thus 
nearly filled, an oyster was gently placed on a glass tripod, where it would rest near the 
center of the desiccator. With clams it was found necessary to leave them in the 
desiccator submerged in water for a few hours before the experiment, because, unlike 
oysters, they would not open quickly after they had been handled. In either case, 
though, water from the shelf reservoir was siphoned into the overflowing desiccator for 
sufficient time to bring the oxygen content to approximate constancy and then the 
cover was put on, leaving just opening enough to admit the siphon tube. 
The side cock of the desiccator was then connected by rubber tubing to the sampling 
bottle, and after starting by suction the water was allowed to siphon through the bottle 
two to three minutes. This period was found quite sufficient to give reliable duplicate 
results for water containing any percentage of oxygen measured in these experiments. 
The water running into the desiccator meantime was in excess of that running into the 
bottle, so that an overflow was maintained from the top of the desiccator and the sample 
rendered as fair as possible. At the exact second recorded as the beginning of the experi- 
ment the side stopcock was closed, the siphon quickly withdrawn from the desiccator, 
From aquarium of running sea water, a part of the aquarium system of the laboratory. 
