64 SENSE OF SMELL IN DOGFISH. 



by the walls of the pool and partly by chicken wire attached to poles driven into the 

 bed. It was freely open to the changes of tide. Fish from an adjacent pool were 

 transferred to it from time to time for experiment and observation. 



When three or four dogfish were liberated in the pen, they swam leisurely up and 

 down its length in a mid depth of water, occasionally resting on the bottom. If, during 

 their swimming, some crushed crabs wrapped in cheesecloth were thrown into the pen, 

 they quickly changed their method of locomotion. When a fish in the course of its 

 ordinary swimming approached to within a few feet of the packet of crab meat, it usually 

 made a sudden movement to one side with its head, swam at once to the bottom, and 

 in quick circuitous turns, often in the form of a figure eight, swept the bed of the pen. 

 In a short time it narrowed its search to the immediate vicinity of the packet and when 

 at last its mouth was brought close to the bait, this was seized, shaken, and carried ofi'. 

 Occasionally the packet was not found and the fish then resumed its ordinary method 

 of locomotion until again by accident it came close to the packet, when it would repeat 

 the movements already described. The crab meat used in these experiments was 

 always wrapped in cheesecloth to exclude the possibility of the recognition of the meat 

 by sight. The dogfish, as is well known (Sheldon, 191 1), has extremely poor vision, 

 at least in daylight, and seldom responds to an object until the latter is within a foot 

 of it. 



If the nostrils of a set of fishes to be experimented upon are filled with cotton wool 

 before the animals are liberated in the pen, no attention whatever is paid to the packets 

 of crab meat, though the characteristic reactions return shortly after the removal of the 

 cotton. These experiments, first tried by Sheldon (191 1), show that the dogfish is 

 excited by food through the stimulation of its nasal organs and that it subsequently 

 locates the food by the same means; in other words, the dogfish scents its food as air- 

 inhabiting vertebrates do. 



The question that I set to decide was that of the precise method by which the dog- 

 fish reached its food. After the initial excitement, does the fish move aimlessly about 

 till, by accident, it runs upon the packet of crab meat, or is it directed toward this packet 

 by some special form of stimulation? In working on this problem it proved important 

 to observe two conditions. First, it was found that dogfish of medium size were more 

 favorable for these experiments than larger or smaller ones. In general, fishes about 2 

 feet in length were found to be most satisfactory, and these were used in most of the 

 experiments. Secondly, the dogfish reacted well only when in companies. A single 

 dogfish in the pen seemed to suffer an inhibition of its activities excepting those directed 

 toward returning it to the school. It swam about with incessant efforts at escape and 

 paid very little if any attention to baits or food. It was therefore found best to work 

 with several fish at once, and usually a set was chosen in which the individuals differed 

 enough in size and markings to make them severally distinguishable. Often in a given 

 set the larger ones were prepared in such a way that they showed no response to the 

 packet of crab meat, and the smallest one of all was made the subject of the special experi- 

 ment. This one individual was provided with a favorable environment for its own reac- 

 tions without being subjected to disturbances from others. 



If the substances emanating from the packet of crab meat exert a directive influence 

 on the dogfish through the sense of smell, evidence of this might be discovered by inter- 



