the: sense op smell in fishes. 
41 
Three of the dogfish, all of which had eaten readily before the use of the cotton, 
but which had refused to do so thereafter, although tested for three successive days, 
were removed from the pool and the cotton withdrawn from the nostrils. These were 
returned to the pool, which now contained no normal fishes. The following day one of 
these ate readily in the usual manner, although there seemed to be slightly more diffi- 
culty than usual in finding the crabs, both in the first place and after one had been 
dropped. A day or two later all three ate as usual. This experiment was repeated 
twice with the same results. 
These experiments indicate that the dogfish normally recognize the proximity and 
location of food through the use of the olfactory apparatus. It may be argued, of 
course, that the mere presence of the cotton in the nostrils renders the fish so uncom- 
fortable that it refuses to eat, even though it act otherwise in a perfectly normal man- 
ner. To obviate this objection four dogfish were removed and one nostril only stuffed 
with eotton. These four only were now placed in the pool. One of these, within an 
hour thereafter, caught a crab after the usual preliminary procedure, but lost it and 
then seemed to take no further interest in the matter. The following day all four ate 
as usual. It was noticed in this experiment, also, that the dogfish had rather more 
diffieulty than normally in finding the food, but that this wore off in a eouple of days. 
These four fish were removed after four days, and seven others, fixed in the same way, 
were substituted. The results were the same. These tests show that the presence of 
the cotton is not sufficiently irritating to interfere seriously with the normal feeding 
habits of the dogfish. As remarked earlier, individuals with the nostrils plugged act, 
except in so far as the feeding habits are concerned, just as do the normal fishes. As 
both kinds swim about the pool, such could not be identified by an uninformed observer. 
The nasal apparatus of the dogfish consists of a pair of large capsules, partially 
divided into two parts by means of a superficial and inclosed flap of skin rostrally and 
a fleshy ridge caudally. There are thus two incompletely separated apertures, a rostro- 
lateral and a caudomedian, the latter closer to the mouth. (See Sheldon, 1909, fig. 3.) 
The capsules contain a double row of lamellae, extending laterally from a median ridge 
much as do the barbs from the rachis of a feather. This ridge extends from the more 
lateral to the more medial opening. The lamellae are innervated by an enormous num- 
ber of short olfactory nerve fibers which terminate in the large olfactory bulbs, closely 
apposed to the capsules. The nervus terminalis of Locy also sends a few fibers into the 
lamellae, while the capsules, in general, are innervated for tactile and general chemical 
sensation by the nervus maxillaris trigemini. 
During the ordinary movements of respiration, as water is taken into the mouth, a 
current is, by suction, drawn through the nostrils, entering the more rostral and leaving 
at the more caudal aperture, to be drawn farther, to some extent, into the mouth. 
This may be easily demonstrated by fastening a dogfish on its dorsum and expelling, 
from a pipette, a colored solution rostrally of the nostrils. The current, then, follows 
the median ridge, a part being diverted laterad between the lamellae. The shape and 
