58 RALPH EDWARD SHELDON 



In all the above experiments a number of different sets of 

 dogfish were used. There were usually from six to eight fish in 

 the pool at a time. Usually only one or two experiments a day 

 were performed in order that there should be no interference 

 between them. 



These observations show, beyond doubt, that the dogfish ob- 

 tains its food through the use of a chemical sense. Experiments 

 were now undertaken to find out what part the olfactory appara- 

 tus plays in these reactions. 



Experiments on fish with occluded nasal apertures 



Four dogfish which had eaten readily when in the normal con- 

 dition, were removed from the pool and their nostrils stuffed 

 with cotton wool ; in two of the cases the cotton was covered with 

 vaseline. When returned to the pool such fishes rush about 

 violently for a few minutes, as do all dogfish which have been 

 out of water. They soon, however, quiet down and swim about 

 the pool as do the normal fish. Twenty -four hours later three crabs 

 were placed, an hour apart, in the pool,, which now contained, in 

 addition, four normal fishes. All were found, in the usual manner 

 and length of time, by the fishes without cotton in the nostrils. 

 At no time did any of the individuals with the nostrils filled show 

 the slightest interest in the crabs, although such often swam with- 

 in a few inches of the food. Moreover, these fish made no attempt 

 to follow those which had secured one of the crabs, although the 

 food was occasionally dropped. It was often observed that two 

 dogfish, one normal and the other with the nostrils filled, would 

 be swimming along the wall side by side when they approached 

 the vicinity of the crab ; the normal fish would then make the usual 

 sudden turn to search for food while the individual with the cot- 

 ton continued on its way with no change in the lazy swimming 

 movement. As it was noted that vaseline or a close packing of 

 the nostrils with cotton caused suppuration in time, this experi- 

 ment was repeated three times with cotton loosely packed. Re- 

 sults, similar to the above, were secured in all cases. Two tests 



