FOOD OF THE OYSTER, CLAM, AND RIBBED MUSSEL. 383 
in an Erlenmeyer flask with a little of the pelagic catch added, in 
order to give it the necessary elements for the growth of the diatoms. 
The flask was then plugged with cotton and sterilized by boiling. 
Afterwards, when it had cooled, a few drops of the pelagic catch con- 
taining but a few diatoms were introduced into this sterilized medium. 
After some days small colonies of diatoms appeared on the wall of the 
‘flask, especially on the side turned toward the light. One of these 
colonies was removed by means of a sterilized platinum needle and 
‘introduced into another Erlemeyer flask containing the same medium. 
‘This culture was afterwards used for experiments. 
The experiments were carried: on in the following way: A hash of 
‘fish and one of shrimps was suspended in water, the suspension con- 
‘taining particles not larger than a copepod. Clams were first used 
‘for the experiments. A culture of diatoms in sufficient quantity to 
‘eause a small, well-defined cloud in the water was offered to them by 
‘means of a fine glass tube, the end of which was brought close to the 
| : : : 
ventral opening of the siphon, care being taken (see figure’ not to 
nt 
itouch it. The culture was now allowed to flow through, and soon 
(disappeared in the opening of the siphon. Many such cultures were 
‘accepted by the clam, but when similar experiments were made with a 
‘hash of fish, the result was either that the opening closed as soon as the 
particles of fish touched it, or the suspension was accepted as before, 
‘but almost as soon as taken it was forcibly ejected and often thrown 
to a distance of six or seven inches. The shrimp hash was rejected in 
the same manner. 
When soft clams were used, the same results were obtained, and when 
the hash was brought between the open shells of the oyster, the same 
phenomena were observed, the suspension being rejected and the shells 
immediately closed, while the diatoms were readily accepted. Though 
these experiments were repeated over and over again, I always obtained 
the same results with a single exception in the case of a soft clam, an 
individual apparently without a discriminating taste, which accepted 
a great quantity of the hash, but finally rejected it also. 
> 
