222 SMELL, TASTE, AND CHEMICAL SENSE IN VERTEBRATES. 



York They were the young of a species of Petromyzon, probably land-locked, 

 which at this stage could not be specifically identified (Gage, 1893, p. 456). 

 They were extremely hardy and satisfactory for experimental work and I am 

 under great obligations to Professor Gage not only for having called my attention 

 to this animal but also for having provided me with a generous supply of living 



In testing the Ammoccetes, they were placed individually in large, clean, 



glass dishes filled with tap-water and kept in a situation where the light was not 



bright, for these fishes are quite sensitive to illumination (Parker, 1905). After 



an individual had come to rest on the bottom of the dish, a pipette-full of distilled 



water was slowly discharged on a particular part of its body. If this operation 



called forth no movement on the part of the fish, as was usually the case, it was 



followed by a similar slow discharge on the same spot of the solution to be tested 



and the absence or nature of the response of the fish was then noted. The 



Ammoccetes was then rinsed in clean tap-water, transferred to another dish, and 



allowed to remain quiet for a period of not less than ten minutes before it was 



subjected to another test. Three regions on the body of the fish were selected 



for these tests, the mouth, the mid-trunk, and the tail. The rapid spread of the 



solution in the surrounding water prevented a more accurate localization. 



The materials that were used for making the test solutions were such as have 

 been frequently employed in testing the sense of taste in man : for a sour stimulus, 

 hydrochloric acid was used; for an alkaline one, sodic hydrate; for a salty one, 

 sodic chloride; for a bitter one, quinine hydrochloride; and for a sweet one, cane 

 sugar. All the solutions were made in distiUed water to which the Ammoccetes 

 did not react; the acid, alkali, and salt solutions were made up as normal solutions 

 and the quinine and sugar solutions as molecular solutions. Although all reason- 

 able care was taken in the preparation of these solutions, it is obvious that tne 

 method of application of them— the discharge of a small amount of solution on a 

 fish in a large volume of tap-water— was such that the solution would suffer a con- 

 siderable and irregular dilution at the moment of application, so that the recora 

 obtained from the tests apply undoubtedly to weaker solutions than those r 

 corded in the tables which really represent the strength of the solutions when tn y 

 were in the pipettes. Notwithstanding this source of error, the r ^P on f es 

 the individual fishes to stimuli assumed to be similar, were gratifymgly uniio 

 The results of these tests are summarized in Table I. 



Solutions of hydrochloric acid of normal and decinormal concentration ^ 

 stimulating to the mouth, mid-trunk, and tail of Ammoccetes, the three pa 

 the body tested. At dilution n/40, the mid-trunk was no longer stimu i , 

 though the mouth and tail were so. At n/80 the tail was not stimulate a, tn ^ 

 the mouth was, and it was not until a dilution of n/2,560 was reached oa 

 mouth ceased to be stimulated. These records show that, though the wh ° engitive 

 surface of the fish is stimulated by hydrochloric acid, the tail is more ^ 



than the mid-trunk, and the mouth vastly more so than the tail s 





