38 
bulletin of the bureau of fisheries. 
Bateson, Nagel, and others beheved that selachians recognize their food through the 
sense of smell ; their evidence is, however, valueless in this connection owing to the fact 
that irritating substances were used as tests, or else no distinction between smell and 
other chemical senses was made. Mr. Vinal Edwards, collector for the Woods Hole 
Laboratory of the Bureau of Fisheries, states that it is the custom in fishing for dogfish 
to throw out in the tide lines baited with menhaden or alewives. For a time no dog- 
fish will be seen, then they will appear in numbers, swimming around the bait in gradually 
diminishing circles until finally it is seized. Field (1907) finds that the dogfish care- 
fully search the bottom for crabs. Finding one they turn on their sides to seize it, then 
dart off quickly, shaking the crab as a terrier would a rat. After swallowing the food. 
Field states that the dogfish keeps up its active swimming, often returning to the place 
where the crab was found. 
Some experiments on the relation of the olfactory apparatus of the dogfish to its 
feeding habits were undertaken in the summer of 1908, but failed owing to the fact that 
the fishes refused to eat in captivity. These experiments and those of Field show that 
dogfish will not eat if kept in large tanks or even in the large cod cars of the station, even 
though they are kept in captivity to the point of emaciation. In order, therefore, to 
give them a habitat comparable to the normal, a portion of the large observation pool 
of the station was fenced off with meshed wire. This gave a pool 24 feet long, 8 feet 
wide at one end, and 10 feet wide at the other, with normal sea bottom, an irregular 
stone wall on three sides, and a depth of water of from 2 to 8 feet, depending on the por- 
tion of the pool considered and the height of the tide. Tufts of eelgrass, together with 
many other varieties of sessile marine life, grew on the botttom and sides. The pool, 
therefore, fulfilled to a reasonable degree the conditions of normal life, so far as the 
dogfish are concerned. 
The individuals used were those caught in the traps from day to day and placed 
in the pool for a period of 10 days in order to bring about a state of hunger. 
Spider or blue crabs were first offered the dogfish, but were always refused. The 
rock crab. Cancer irroratus, was next tried, with success, and used for all the experimental 
work. All experiments were conducted at low tide, when it was easy to observe the 
actions of the individual fishes. At first living crabs were used. These were found 
by the dogfish in from 10 to 15 nfinutes. Next, crabs were killed and a hole broken 
in the carapace, exposing the flesh. Such were found in from 2 to 5 minutes. These 
results suggest at the start that the food is recognized through the diffusion of animal 
juices into the water. Crabs killed, with the flesh thus exposed, were used for all 
further work. 
In a total of about 40 experiments the method of feeding was the same in all cases. 
The dogfish spent most of the time swimming lazily around the pool, usually close to 
the sides. Now and then the direction was reversed, but at no time was there observed 
any search over the bottom or the rocks forming the sides of the pool. When a crab 
was placed in the pool, a few minutes was required, as noted, before any evidence of 
stimulation was to be seen. Then one of the dogfish which happened to be swimming 
within 3 or 4 feet of the crab seemed suddenly startled. It turned very quickly, and 
