THE SENSE OF SMELL IN FISHES. 
39 
swimming with quick, nervous motions, instead of the calm, lazy movement of the 
unstimulated fish, began a systematic search over the bottom, investigating particularly 
grassy or uneven spots. The head was moved rapidly from side to side as the fish swam 
slowly, coursing, in gradually diminishing circles, 2 or 3 inches from the bottom. 
When within 2 or 3 inches of the crab the dogfish seized it suddenly, making off in a 
swift rush. As remarked by Field, the crab is shaken violently from side to side for a 
moment, as the shell is crunched and broken by the powerful jaws of the fish, after 
which it is quickly swallowed. Occasionally, however, the crab is dropped during the 
process; when this occurs, a search similar to the first follows until it is found again. 
At no time did the dogfish appear to make any use of the sense of sight in feeding. 
A crab hidden in eelgrass is found as quickly as one lying on the open bottom; 
moreover, one is found with equal promptness whether lying on its venter, exposing 
the dark carapace, or on its dorsum, with the light-colored venter showing conspicu- 
ously. A dogfish, dropping a crab, is apparently unable to find it again excepting by 
means of the same sense which enabled recognition of food in the first place. It was 
observed, however, that a dogfish with food is usually followed by others in the vicinity, 
which endeavor to secure possession of it. Moreover, the fish will frequent for some 
time thereafter the region of the pool in which food was found. This is probably due 
to olfactory stimuli, although sight may be brought into play to a slight extent. It 
was often noted that a dogfish would circle around the spot where a crab had lain, often 
biting into the bottom at the exact spot; probably some body juices had escaped into 
the ground. Now and then a crab was placed on the bottom near the screen separating 
the experimental from the larger pool. If no fish capable of finding food were present 
in the former, it would often happen in a few minutes that 8 or 10 dogfish from the 
large pool would be swimming rapidly back and forth along the screen, endeavoring 
to find their way through. 
Some experiments were tried with a hook and line baited with the flesh of a rock 
crab tied in cheesecloth. The dogfish here follow the same procedure as when the crab 
lies on the bottom. On noting the proximity of food the fish begin to swim, as before, 
in circles, but for a time persistently search the bottom beneath the baited hook. At 
length they gradually rise, turning somewhat sidewise, as stated by Field, to seize the 
bait. When the food is lying on the bottom this sidewise turning was never observed 
Observation of the feeding habits of the dogfish would indicate, then, that it recog- 
nizes its food through some chemical sense. To test this the following experiments 
were performed: 
Some fresh eelgrass was secured and two packets, closely resembling each other, were 
made, one containing a small stone, while the other inclosed a crab. Both were so tied 
that when placed in the water a foot apart a portion of the grass rose toward the surface, 
giving an appearance similar to the grass of the pool. In three sets of experiments the 
presence of food was detected in an average time of three minutes, the packet found by 
the usual procedure, torn apart, and the crab eaten. At no time did the packet contain- 
ing the stone receive the slightest attention. 
