THE SENSE OF SMELL IN BIRDS 637 
one of the four chambers (fig. C, f.b.), and the odor bottle was 
connected with the chamber which contained the food, where 
it replaced the bottle located there in the previous series of experi- 
ments (fig. A, 6’.). Newspapers were laid underneath the entire 
labyrinth to prevent the birds from soiling the floor. 
After the apparatus had been inspected to insure its being ready, 
the air pumps were started by turning on water at P, fig. B. 
The cage containing the bird was placed with the door against 
the entrance to the enclosure (fig. A, d); the cage door was then 
raised, allowing the subject to enter. It was not found practi- 
cable, though desirable, to place this entrance at the center be- 
cause there was greater need of room for the exhaust funnel 
at that point. Immediately after the dove was admitted, air 
pressure was started by turning a cock at s.c., fig. D. This order 
of procedure insured the entrance of the odor into the enclosure 
by the time the bird started to look for food, and the postponement 
of turning on the air pressure until just after the bird had entered 
decreased the danger of a general diffusion of the odor in the cen- 
tral enclosure during the time which was ordinarily occupied by 
the subject in hunting for the food. 
When the bird reached the seed, the air pressure was cut off 
to avoid unnecessary diffusion of the odor, but the exhaust was 
kept going until the last of the birds used in each set of experi- 
ments had found its food. 
When the last bird had finished feeding and had been removed 
to its cage, the entire labyrinth was elevated several inches above 
the floor for ventilation ; a window and a door to the room contain- 
ing the apparatus were kept open for a short time at least during 
the ventilation. 
In order to determine so far as possible. the extent of the space 
in the main enclosure where the odors which were employed in 
the experiments might be expected to be perceptible, various tests 
were made. Strips of litmus paper about eight inches long were 
hung from standards placed at thirty-six approximately equal 
points inside the main enclosure. These strips were all dipped in 
distilled water just before air pressure was turned on. Ammonia 
water of several dilutions was placed in a bottle which was used 
JOURNAL OF MORPHOLOGY, VOL. 22, NO. 3 
