THE ORGAN AND SENSE OF TASTE IN FISHES. 
257 
the meat, but not always nor so promptly as with fresh meat. Upon testing the 
sides of the body again after this experience I got a reaction. The fishes would turn 
and touch the meat with the barblet or lips before taking it, rarely giving the quick 
reaction characteristic of fresh meat. Evidently the cooked meat has less taste to 
the fishes than fresh meat and this interferes with the reaction. They eat the cooked 
meat when they are sure that it is edible. 
These experiments, all of which were many times repeated and controlled, I 
think show conclusively that practically the whole cutaneous surface of Ameiurus is 
sensitive to both tactile and gustatory stimuli, and that the latter call forth charac- 
teristic reflexes which are of the greatest value to the fish in procuring food. The 
fish normally reacts to contacts on the body by both types of stimuli — to the mere 
tactile stimulus (if at all) by a tentative movement calculated to bring the doubtful 
substance into contact with the more highly sensitive barblets or lips, but to the 
tactile stimulus accompanied by the gustatory by an immediate, rapid, and precise 
movement calculated to seize the food. This latter reflex is unvarying and is very 
persistent under a great variety of forms of stimulation. The former (“ tactile”) 
reflex is less stable, and may be readily eliminated by a simple course of training. 
Clearly the gustatory element of the sensation complex resulting from a contact with 
a sapid substance is more important than the tactile element. 
It is clear that in order to call forth the characteristic “gustatory” reflex the 
stimulus must be quite strong* and rather sharply localized. For when there is only 
a small amount of meat juice diffused through the water, as by the presence of a 
piece of fresh meat near the fish, he is not able to localize it accurately, but exhibits 
only the “seeking reaction.” I have not as yet been able to convince myself whether 
the fish could accurately localize a strong and sharply localized gustatory stimulus 
with no tactile element. In all the experiments in which meat juice was directed 
against the body with a pipette or syringe there was doubtless some tactile effect 
produced by the impact of the jet. We know from the experiments that pure tactile 
stimuli can be accurately localized on the skin, and there can be no doubt that under 
normal conditions these assist in the localization of the food object. Compare the 
further discussion in the Addendum, pages 270-271. 
EXPERIMENTS ON GADOID FISHES. 
The preceding experiments were all carried on in the zoological laboratory of 
Denison University; the experiments on marine fishes which follow were made during 
the summer of 1902 at the U. S. Fish Commission laboratory at Woods Hole'. The 
feeding reactions of three types of gadoids were studied, viz, young pollock ( Polla - 
chius virens ), about 10 cm. long; hake ( JJrophycis ten uis), about 20 cm. long, and 
young adult tomeod {Microgadus tomcod). 
As is well known, the hake and tomcod have a mental barblet which ds known to 
be abundantly set with terminal buds and which receives both communis and general 
cutaneous innervation. In all three types the lips are freely supplied with terminal 
buds and there is a recurrent branch of the facial nerve, the ramus lateralis acces- 
sorius, which carries communis fibers into the trunk to supply terminal buds found 
on the fins, especially the free rays of the ventral or pelvic fins. These fins are far 
forward under the throat. In the pollock they are but little modified; in the tomcod 
F. C. B. 1902—17 
