FORAGING BY OLFACTION IN THE GRAY JAY
, Ronald G. Langley
B.A., University of Colorado at Denver, 1987
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Arts
Department of Biology
This thesis for the Master of Arts degree by
Ronald G. Langley
has been approved for the
Department of Biology
Diana F. Tomback
Langley, Ronald Glenn (M.A., Biology)
Foraging by Olfaction in the Gray Jay
Thesis directed by Associate Professor
Diana F. Tomback-
Olfactory ability in birds, and especially in
passerines, has been considered minimal based on their
olfactory bulb/cerebral hemisphere ratios. Recent
observations have indicated that a reexamination of
this assumption is needed. Gray jays were tested to
determine whether they use olfactory cues during
foraging. Prior to experimentation, each jay was
trained to find food in one of two artificial trees.
In a series of trials, items of rancid or fresh ground
beef, processed cheese or dried bread were placed at
random locations in a tree with 30 or 52 possible
sites, and all possible search locations were covered
by paper cups, eliminating visual cues. Food deprived
jays searched until the first hidden food item was
found. The ability of jays to locate rancid ground
beef was significantly better than random foraging
according to a null model (p<0.001) and their ability
to locate the dried bread controls (p=0.018). The
performance of jays searching for fresh ground beef
improved after a period of sensitization, and one jay
found processed cheese by olfaction.
Olfactory bulb/cerebral hemisphere ratio
percentages were obtained from 4 gray jays and from 1
or 2 specimens of each of 5 other corvid species. The
ratio percentage obtained for gray jays is 7.8%,
similar to the mean ratio percentage of 7.2% that was
obtained from all the corvid specimens combined with
information on corvids from the literature.
Results of the experiments indicate that gray
jays may use olfaction to some extent while foraging
for food, particularly carrion which is an important
food for this species. This study suggests that the
role of olfaction in foraging by some passerines has
been underestimated by previous investigators. Small
olfactory bulb/cerebral hemisphere ratio percentages
may not rule out specialized olfactory ability. When
testing for the use of olfactory cues vs. other kinds
of cues by foraging birds, investigators must use
natural food items or their results could be
The form and content of this abstract are approved. I
recommend its publication.
Faculty member in charge of thesis
The completion of this study would not have
been possible without the assistance of several
people. Lorraine Reiner, of the University of Denver
High Altitude Lab, allowed me to collect jays at her
banding station and donated specimens for anatomical
work. Jim Ha, at Colorado State University, and Town
Peterson, from the Field Museum of Natural History,
also contributed specimens for this study. The Denver
Museum of Natural History donated specimens for
practice measurements of the brain. Janis Driscoll
was very helpful with statistical advice and support
throughout my graduate years. Kristi Burnell
collected data and assisted with development of the
experimental design. I am especially grateful to
Diana Tomback, my advisor, for her continual
assistance and patience during this study.
I. INTRODUCTION .............................. 1
Purpose of the Study.................... 1
Natural History of the gray jay. ... 5
II. MATERIAL AND METHODS...................... 7
Capture and maintenance of jays. ... 7
General experimental procedure .... 7
Training ............................... 8
Experimental Set-up #1 ................ 10
Experimental Set-up #2 ................ 11
Null Mathematical Model.................16
Anatomical Studies .................... 17
Null Mathematical Model.................18
Foraging Experiments .................. 18
Anatomical Results .................... 26
1. Null mathematical model of random
2. Number of cups removed by jays searching
for rancid ground beef........................21
3. Number of cups removed by jays searching
for dried bread...............................23
4. Number of cups removed by jays searching
for fresh ground beef.........................27
5. Number of cups removed by jays searching
for processed cheese..........................28
6. Olfactory bulb/cerebral hemisphere
ratio percentages for several corvid
1. Mean performance of jays searching for
rancid ground beef and dried bread. .
Purpose of the Study
Birds as a group have been considered to
possess poor olfactory ability (Walker 1987).
Experiments in recent decades have challenged this
generalization for some non-passerine birds.
Olfaction has been shown to be used by some species
for long distance navigation, finding nesting cavities
and foraging. In a classic study by Stager (1964),
turkey vultures (Cathartes aura) were able to detect
the smell of rotting meat. More recently, turkey
vultures were shown to react to the constituent odors
of rancid beef (Smith and Paselk 1986).
Procellariiforms have the ability to search for food
by olfaction (Hutchison and Wenzel 1980). Leach's
storm petrel (Oceanodroma leucorhoa) has been reported
to find its nesting burrow by olfaction (Grubb 1974).
Navigation in pigeons (Columba livia) is thought to be
guided, at least in part, by olfaction (Schmidt-Koenig
1987, Waldvogel 1987)
A majority of birds, primarily the
Passeriformes, are still considered to possess poor
olfactory ability; this belief stems from a low
olfactory bulb/cerebral hemisphere (OB/CH) ratio
percentage for many members of this order (Bang and
Cobb 1968). Birds with a ratio percentage greater
than 18%, which is the lower end of scent
discrimination (Bang 1971), are probably able to use
their olfactory sense in various activities Of the
previously mentioned species, the turkey vulture has a
ratio percentage of 27%, Procellariiforms 27.8%,
Leach's storm petrel 29.4% and the pigeon 18% (Bang
and Cobb 1968, Bang 1971).
Species of the family Corvidae (Order:
Passeriformes) have one of the smallest OB/CH ratio
percentages of all bird speciesbetween 6 and 8%
(Bang and Cobb 1968). Studies of corvid olfaction by
several experimenters have suggested that these birds
do not detect food substances by olfaction. Bossema
(1979) determined that the European jay (Garrulus
qlandarius) could not find peanuts, acorns or cheese
by olfaction. Baida (1980) and Vander Wall (1982)
working independently with the Eurasian (Nucifraqa
carvocatactes) and Clark's nutcracker (N. Columbiana),
respectively, found that these birds could not recover
caches of pine seeds by olfaction but instead used
spatial memory. The gray jay (Perisoreus canadensis)
in a study by Bunch and Tomback (1986) could not
locate hidden processed cheese or pungent smelling
beer cheese by olfaction.
In contrast, studies within the past few years
on two corvid species have shown olfactory ability.
Harriman and Berger (1987) demonstrated that ravens
(Corvus corax) could find buried fish and canned cat
food. Black-billed magpies (Pica pica), studied by
Buitron and Nuechterlein (1985), were able to find
hidden suet or raisins coated with cod-liver oil at a
greater frequency than uncoated caches. Other
observations by Buitron and Nuechterlein (1985)
indicated that magpies could smell rotting chicken
Electrophysiological experiments indicate that
birds with OB/CH ratio percentages as low as 3% can
respond to odorants (Bang and Wenzel 1985). Peterson
(1977) determined that the breathing and heartrate of
black-billed magpies increased when the odor of
carrion was presented. Thus, it is possible that
corvids and other passerines can respond to and use
olfactory cues to forage for some food items.
Although Bunch and Tomback (1986) determined
that the gray jay could not find cheese by olfaction,
it is possible that this corvid, despite a small OB/CH
percentage ratio, can detect a few ecologically
relevant scents. Gray jays eat meat as part of their
natural diet, and they may be able to find this food
by using olfaction. In this study, I test the
hypothesis that gray jays, despite a low OB/CH ratio
percentage, can detect hidden carrion (fresh killed or
rancid meat) by olfaction.
NATURAL HISTORY OF THE GRAY JAY
The gray jay inhabits coniferous forests in
Canada, with its range extending into the Cascade and
Rocky Mountains in the U.S. (Robbins et. al. 1983).
This bird has also been reported to have invaded
southern Minnesota (Eckert 1987, Janssen 1987). The
gray jay is territorial year-round and defends an area
one to four km2 (Rutter 1969). Like other corvid
species (Goodwin 1986), the gray jay caches food items
but primarily during fall and winter (Rutter 1969).
Food items are formed into a bolus with the tongue and
stored in branches, bark crevices and foliage (Rutter
Gray jays, have a diverse natural diet composed
of 4 main food types: 1) vegetable matter (primarily
berries), 2) eggs and young nestlings, 3) insects and
4) animal carcasses of all types (fish, reptile, bird,
mammal) (Rutter 1969, Ouellet 1970, Goodwin 1986).
Animal meat may be an important food source during
winter months for the gray jay; Harper (1953) and Dow
(1965) suggested that winter survival may be enhanced
by the use of animal carcasses. In northern areas,
carcasses are primarily provided by predator kills
(Dow 1965); however, in my study population, meat is
usually obtained from road kills and human hunting
activities. Survival might be further enhanced if the
gray jay could locate these sources of meat by
MATERIALS AND METHODS
Capture and Maintenance of Jays
Gray jays were captured with treadle-style
walk-in traps, baited with bread, in the vicinity of
Echo Lake, Clear Creek County, Colorado. Trapping was
conducted during fall and early spring from 1985 to
1988. Two to six jays were housed at one time in
individual cages (Prevue Hendryx) measuring 66 cm x 66
cm x 99 cm (height) in the Avian Ecology Laboratory
(1985 1987) or the Animal Care Facility (1988) at
the University of Colorado at Denver. Jays were fed a
daily diet of sunflower seeds, dog food (Gaines
burgers), bread and processed cheese. A natural
light-dark cycle was simulated during jay captivity.
Whether or not a jay was a successful subject, each
jay was banded with a U.S. Fish and Wildlife band and
released at the trapping site. Recaptured jays were
not used again.
General experimental procedure
Experiments were conducted in an indoor aviary
measuring 2 m x 2 m x 2m (height), with a one-way
glass window for observation. Inside the aviary was a
tree constructed of wood dowels, with 3 or 4 pinyon
pine (Pinus edulis) cones on each branch (see
experimental set-up #1 and #2). Prior to an
experimental trial, the observer hid food in 3 to 5
cones, selected by random numbers table, and covered
all cones with paper cups (Dixie 2.5 oz) so that the
food was not visible, even when a jay looked under the
cone. Ten pin-holes were punched in the bottom of
each cup to permit odor to escape. When a jay was
introduced into the aviary, it searched for the hidden
food by knocking cups off the cones until a food item
was found. The number of search attempts, i.e. cup
removals, were recorded. Trials were not counted when
jays did not search for food or knocked cups off cones
in a hyper-active fashion (see Motivation).
Nine jays completed training and experimental
trials. Jays were assigned a number according to
experiment set-up used and the particular jay tested,
e.g., 2-5 signifies set-up #2, jay 5.
Gray jays were trained in their cages to
remove cups covering pinyon pine cones and retrieve
food as follows. Two pine cones were attached to a
perch in each cage, and jays learned to remove food
from these cones. When the jays were accustomed to
finding food in the cones, paper cups were placed over
the cones to hide the food items. The jays learned to
remove the cups to find the food. Initially, cups
were placed over the cones loosely. The attachment of
cups was made firmer with tape placed on the inside of
the cups, so that jays had to pull a cup to remove it.
Most jays learned to easily uncover a cone, even with
the tape adjustment.
Each jay was habituated to its cage for 2 to 3
days and then placed in the aviary for periods of 1 to
2 hours without food on 3 successive days. Jays were
then placed in the aviary with several cheese pieces
(approx. 0.5 cm3 each). The jays formed boluses from
the cheese and hid them (cache) in the pine cones.
These caches were recovered by the jays the following
Most jays required food deprivation before
they would cache. If the birds did not cache, they
were deprived of food in their cages from 2 to 8 hours
before being placed in the aviary again. Caching
periods in the aviary ranged from 15 min to 1 hour.
Jays were returned to their cages after several caches
were made or until 1 hour had passed. Food was
returned to the jays in their cages at random time
intervals from 0 to 3 hours. If the jays immediately
received food upon returning to their cages, they
would stop caching and recovering in the aviary. Jays
that did not learn to cache or recover were released
at the capture location.
Food deprived jays were allowed to recover
their own caches. Jays were given 3 cache and
recovery trials without cups on the cones. Cups were
then placed over all the cones on successive trials,
requiring the jays to recover their caches without
visual cues. Training trials were administered for
each jay until the number of search attempts was
better than chance expectation based on a null model
of random searching behavior (see Null Model). A
minimum of 2 trials was given. If trained properly,
most birds performed well on both trials.
Experimental trials were then started. After
experimentation was complete, all jays were returned
to the trapping location.
Experimental Set-up #1
The experimental tree for this set-up was the
same as that used by Bunch and Tomback (1986) (see
Bunch (1985). for details). The tree had 13 randomly
distributed branches, with 4 equally spaced pinyon
pine cones on each branch. Thus, 52 search sites,
i.e., 52 cones, were available. Each site was
assigned a location number. Five pieces of rancid
ground beef (approx. 0.5 cm3) were placed by the
experimenter in randomly selected cones (location
chosen from random numbers table). All cones on the
tree were covered with paper cups.
Each bird was deprived of food for a length of
time designed to optimize motivation for searching
(see Motivation). Cup removal or cups tilted by jays
greater than 30o were recorded until a food item was
found. The jays were then allowed to eat the food
item and the trial was ended. A jay had to complete 5
rancid beef trials for its data to be counted. Beef
was set out for 3 to 5 days at room temperature to
spoil before use. Three jays were given 5 rancid beef
trials and one jay was given 2 fresh ground beef
Experimental Set-up #2
Procedures followed that of set-up #1 with 3
exceptions: 1) a different artificial tree was used,
2) punishment was given for searching errors and 3)
more food types were presented.
The artificial tree was changed to facilitate
training jays. Fifty percent of the jays captured
could not learn the task or apparently became confused
by too many search sites close together. To alleviate
possible confusion of the jays, another artificial
tree was constructed with only 10 branches and 3
evenly spaced cones on each branch. Thirty possible
search sites were therefore available.
The tree consisted of one 4 cm (diameter) pole
spanning the height of the aviary. Each branch was a
61 cm long, 1.6 cm (diameter) dowel, positioned
randomly on the tree with two restrictions:
1) Neighboring branches were separated by a minimum of
30o to eliminate the problem of odor from one cup
interfering with another cup immediately below or
above. 2) Branches had to be separated by minimum of
4 cm and no greater than 9 cm for structural support.
Branches started at 32 cm above the tree base.
Punishment was administered for an incorrect
decision. The investigator turned off the aviary
light for 10 seconds for the first mistake. This was
doubled for each additional mistake. Punishment was
used to stop birds from hyper-active behavior, i.e.,
rapid methodical removal of all cups, which
occasionally occurred in set-up #1. Cup removals by
jays were recorded until a mistake was made after a
food item was found, or 3 minutes had elapsed after a
food item was located. Occasionally, a bird would
find a second food item immediately after the first.
This information was not available in set-up #1, since
jays were removed immediately after finding a food
Food items presented in this experiment, in
order of presentation, were rancid ground beef, fresh
ground beef, cheese and dried bread. Rancid beef was
prepared by warming 5 g of ground beef at 38o c for 24
hours. Dried bread was selected because, according to
human olfactory ability, it has the least scent of all
food items. If jays have similar perceptions, the
bread trials controlled for visual cues. If birds
found this item as easily as the other foods, visual
cues may somehow be available. Trials with one food
type were completed before another food was presented.
Not all jays received each food type.
Initially, a few preliminary trials were
conducted with random presentation of food items. The
searching behavior of birds was inconsistent with this
regime, and one jay stopped searching altogether.
Sequential presentation also allowed me to determine
whether the jays became sensitized to the scent of
food items and improved their performance over trials.
Three items of each food type were placed in
random locations in the tree (cone location determined
by random numbers table); however, they could not be
on the same branch. This restriction was made to
disperse the odor in the tree. Food was never located
in cones used on the previous trial. Jays had to
complete 5 trials, as in set-up #1, with rancid beef
for the data to be counted. Six jays completed this 5
trial criterion. Jay 2-1 had 3 trials and jays 2-2
and 2-3 each received 5 trials with fresh ground beef.
Three jays were given processed cheese trials; jay 2-3
was given 3 trials, and the other 2 jays each received
5 trials. Five dried bread trials each were
administered to jays 2-2 through 2-6. After each
trial, all cones that contained food items and 3
unused cones were washed with soap and water to remove
remaining food particles and eliminate residual odor.
All jays were deprived of food before an
experimental trial to motivate them to search for
food. Experimental trials were not used if the jays
did not search properly. When the jays were not
hungry enough, they sat on a branch for a minimum of 3
minutes and sometimes removed nearby cups. When the
birds were too hungry, within 10 seconds of entering
the aviary, they methodically and rapidly removed all
cups encountered. First, the birds removed all cups
on one branch within seconds and then continued to
remove cups on other branches. When this occurred, it
was clear that the jay was not searching in the tree
by olfaction but was simply removing all cups until it
found a food item. If either of these behaviors
occurred, the jays were removed from the aviary and
the experimental trial was attempted the following
I attempted to achieve a state of motivation
between these two extremes, so that birds would move
around the tree searching for food and removing cups.
The motivational state had to be adjusted for each
bird. An average food deprivation time of approx. 2
to 2.5 hours was sufficient for most jays, although
the range of food deprivation was 35 min to approx. 16
hours (overnight). If a jay showed too little
motivation or were too hungry, the experimental trial
was ended and the period of food deprivation increased
or decreased accordingly.
During training, searching behavior was
observed each time a jay was released into the aviary.
If a jay did not show proper motivation, food
deprivation was further adjusted for each jay.
Experienced jays usually required less food
deprivation as they adjusted and learned the
experimental paradigm. When a new food item was
presented, motivation was monitored carefully and food
deprivation adjusted to attain proper searching
behavior; most changes involved an adjustment of 10 to
Food deprivation for all birds was no longer
than necessary to achieve a proper motivational state.
Jays were hungry enough that they would search for
food but not physiologically stressed. All jays
gained a mean of 7 g body mass in the lab and were fed
ad lib in the cages.
Null Mathematical Model
A null mathematical model based on random
searching behavior was devised for each experimental
tree. Random locations (obtained by random numbers
table) corresponding to cone locations were obtained,
taking the restrictions into account. Five random
locations were obtained for tree 1 and 3 locations for
tree 2, corresponding to the number of hidden food
items. A second random numbers table was used to
generate the number of cones searched by a bird
without any cues. Random numbers were generated
without replacement until a number corresponding to
one of the hypothetical food items occurred. The
number of random searches until an item was found
corresponded to one experimental trial. Two hundred
random searching trials were generated for each
experimental tree. The mean, median, standard
deviation and range of search attempts for all 200
trials was calculated for each tree.
Measurements of the cerebral hemisphere and
olfactory bulb were obtained from donated specimens of
gray jays and several other corvid species. None of
my study animals were sacrificed for this analysis.
Frozen specimens were fixed in 10% buffered formalin
for 2 weeks. The specimens were then placed in a
water bath for 24 hours to remove the formalin and
preserved in 70% ethyl alcohol.
The method of measurement followed that of
Bang and Cobb (1968). Longest linear dimensions, to
0.1 mm, of both the olfactory bulb and cerebral
hemisphere were obtained using a calipers.
Measurements were repeated five times for each brain.
A mean was obtained for both the olfactory bulb and
cerebral hemisphere. From these means a ratio
percentage of olfactory bulb to cerebral hemisphere
was calculated for each specimen and a mean ratio
percentage was calculated for each species.
Null Mathematical Model
The null model exemplifies the searching
behavior of a bird in the experimental set-ups without
cues present. Two different models were made, one
model for each experimental tree. The random search
attempts for each model resulted in a median of 7
attempts for tree 1 and 8 attempts for tree 2 before a
food item was found. If all of a jay's search
attempts were below the median for the appropriate
null model, it was assumed that the jays were
searching in a non-random manner. The mean search
performance, obtained from 200 simulated trials for
each tree, was 8.63 and 8.40 for trees 1 and 2,
respectively (Table 1).
Properly motivated jays released into the
aviary immediately began searching for food items.
Jays moved from branch to branch rarely removing a
cup. This occured in the majority of experimental
trials. Experimental trials rarely lasted for more
Table 1. Descriptive statistics
random searching behavior.
Tree #1 8.63
(52 cones with 5
hidden food items)
(30 cones with 3
hidden food items)
from the null mathematical model
Median St. Dev. Min.
7 6.54 1
8 5.41 1
than 5 min when olfactory cues were present, with jays
typically finding a food item within 1 min. After
food was found, the jays ate the food without any
delay for all food items administered, including
bread. Trials in which the jays were not motivated to
search or searched hyperactively, methodically
removing cups off branch after branch (see
Motivation), were not counted.
Nine jays each completed 5 rancid beef trials
(from both experimental set-ups). Due to the small
number of jays used, the results from each
experimental set-up were combined for statistical
calculations. All rancid beef trial performances for
each bird were below the median of the null model for
the appropriate set-up, suggesting that each search
was non-random (Table 2). A repeated measures ANOVA
was computed to determine if jays changed their
performance over trials. This analysis was not
significant, indicating that the jays' performance did
not change systematically over the 5 experimental
trials (F=1.71, df 32,4, p=0.17). The fact that jays
did equally well on the first and last trial indicates
that they were able to detect rancid beef immediately.
To determine if jays performed better than
chance, the search performance for rancid beef was
compared to that of the null mathematical models.
Table 2. Number of cups removed until one item of rancid
ground beef was located.
Jav # 1 2 3 4 5 Mean
1-1 1 1 1 1 1 1.0
1-2 2 2 1 2 2 1.8
1-3 3 3 3 2 3 2.8
2-1 2* 1 1 2 3 1.8
2-2 2 2 1 2 3 2.0
2-3 3 3 3 2* 3* 2.8
2-4 4 2 2 4 1 2.6
2-5 2 1 3 1* 3 2.0
2-6 2 1 2 2 6 2.6
Trial Mean 2.8 1.8 1.9 2.0 2 : .8
* A second first food food item item was was found obtained. immediately after the
Because performance did not change over the 5 trials,
scores were averaged for all trials for each bird.
Because 2 different tree set-ups were used for rancid
beef trials, it was necessary to use two different
models to represent chance performance. Therefore, a
matched sample t test was used to compare each jay's
mean performance to the mean of the appropriate null
model. A significant difference was obtained
(t=29.23, df 8, p<0.001), indicating that the birds
were able to detect rancid beef better than the null
Dried bread was administered to 5 jays in
experimental set-up #2. Changes in performance over
the five trials with bread were evaluated with a
repeated measures ANOVA. The ANOVA indicated that the
jays improved significantly over trials (F=4.5,
df 4,16, p=0.049). This effect can be attributed to a
larger mean number of search attempts during trials 1
and 2 (Table 3).
Using the 3 best trial means (trials 3
through 5), a single-sample t test was used to
determine if the jays were finding dried bread better
than the null model would predict. The jays did not
find dried bread better than random search would
predict (t=0.36, df 4, p>0.05).
Table 3. Number of cups removed until one item of dried bread was located
Jav # 1 2 Trial # 3 4 5 Mean Mean of Trials 3-5
2-2 14 22 7 2 14 11.8 7.7
2-3 14 5 5 15 6 9.0 8.7
2-4 9 5 6 6 3 5.8 5.0
2-5 19 16 6 9 6 11.2 7.0
to i CM 16 6 6 2 2 6.4 3.3
Trial Mean 14.4 10.8 6.0 6.8 6.2
Using the 5 jays that completed trials with
both rancid beef and dried bread, I determined whether
performance was better for rancid beef with a repeated
measures t test. A t of 3.89 was obtained
(df 4, p=0.018), indicating that the jays were able to
locate rancid beef better than the dried bread
Fresh beef and processed cheese trials were
also administered to some of the jays. Because of the
small number of jays that received these trials,
statistical comparisons were not done; however,
general results are reported. Jay 1-3 received 2
fresh beef trials, jays 2-1 three trials, and jays 2-2
and 2-3 each received 5 trials for a total of 15 fresh
beef trials. Five cheese trials were given to each of
jays 2-1 and 2-2, and jay 2-3 received 3 trials for a
total of 13 trials.
When jays were offered fresh ground beef, the
number of search attempts, except for that of jay 2-2,
were fewer than the median of the respective null
models. Two of the five trials of jay 2-2 resulted in
search attempts greater than the median of the null
model; however, improvement over the five trials was
observed. This jay's final fresh beef trial was a
perfect recovery. All jays except 1-3, which had a
perfect recovery on both trials administered, showed
771 Randd Beef |\ \| Dried Bread
Figure 1. Mean performance of gray jays searching for
rancid ground beef and dried bread. Means obtained
from trials 1 through 5 for rancid ground beef and
trials 3 through 5 for dried bread.
improvement over trials with fresh beef. For all the
jays, the performance on trial 5 are below the median
of the respective null models (Table 4).
Processed cheese was presented to 3 jays in
experimental set-up #2. Two of the jays were not able
to find the cheese quickly, and their trial
performances were both above and below the median of
the null model; however, jay 2-3 found the cheese
perfectly on all three trials (Table 5). The results
of jay 2-3 indicate that it was able to home in on the
odor of cheese.
Four gray jays were used for anatomical
measurements, and 1 or 2 specimens each of 5 other
corvid species were also measured (see Table 6). Gray
jays have a mean olfactory bulb size of 1.2 mm and a
cerebral hemisphere measurement of 15.4 mm. A mean
olfactory bulb/cerebral hemisphere ratio percentage of
7.8 was obtained for the gray jays. Of the four jays
measured, the ratio percentage ranged from 4.8 to
10.2%. A mean ratio percentage for corvid species of
7.2 was obtained from my study specimens and from the
Table 4. Number of cups removed until one item of
fresh beef was located.
Jav # 1 2 3 4 5
1-3 1 1
2-1 7 5 5
2-2 15 10 3 4 1
2-3 2 ** * 2 1 1* 1
* A second food item was found immediately after
the first food item was obtained.
** A third food item was found immediately
the second food item was obtained.
Table 5. Number of cups removed until one item of
processed cheese was located.
Jav # 1 2 3 4 5
2-1 18 21 5 2 14
CM 1 CM 5 3 1 13 12
Table 6. Ratio percentages for several species of the family Corvidae.
Species Ratio% N Reference
Blue Jay (Cvanocitta crltata) 6.8 1 Bang and Cobb 1968
Black-billed Magpie (Pica Dica) 7.1 5 Buitron and Neuchterlein 1985, Peterson 1977
Indian Tree Pie (Dendrocitta vacrabunda) 7.0 1 Bang 1971
Common Crow (Corvus brachvrvnchos) 5.0 1 Bang 1971
Gray Jay (Perisoreus canadensis) 7.8 4
Clark's Nutcracker (Nucifracra Columbiana) 7.4 2
Scrub Jay (Aphelocoma coerulescens nevadae) 5.7 2
Scrub Jay (Aohelocoma coerulescens oocleptica) 8.3 2
Pinon Jay (Gvmnorhinus cvanoceohalus) 7.6 1
Yellow-billed Magpie (Pica nutalll) 9.5 1
Mean Ratio = 7.2
The olfactory bulb/cerebral hemisphere ratio
percentage of the gray jay is within the range of
other corvid species. The gray jay has a mean
olfactory bulb/cerebral hemisphere ratio percentage of
7.8, regarded as low for functional olfaction (Bang
1971). Other corvid species have percentages similar
to the gray jay, between 5 and 9%.
Despite this small olfactory bulb/cerebral
hemisphere ratio percentage, gray jays apparently do
have an ability to forage for food by olfaction. All
odoriferous food items presented, in the absence of
other cues, were found by the jays by
olfactionrancid and fresh beef and possibly cheese.
The jays that were tested with rancid ground beef were
able to locate the beef by olfaction; however, when
tested with dried bread, the jays were not able to
find the bread better than chance. The birds were
able to find the rancid beef better than the dried
bread, indicating that they were not using
non-olfactory cues to find any of the food items.
The jays, when tested with dried bread,
improved their performance on the last three trials.
It is believed that the greater number of search
attempts in trials 1 and 2 is related to a changed and
reduced olfactory stimulus. The jays removed more
cups in these trials before adjusting to the new
search conditions. After the birds adjusted to
altered and reduced olfactory cues, they removed fewer
cups, but their search attempts were random.
My population of jays is known to eat meat in
nature from road kills and the remains of hunting
activities (L. Reiner, pers. comm.). One of the jays
was trapped during hunting season with fresh blood on
its bill and claws. Gray jays and other corvid
species at Echo Lake have also been observed eating
the meat from road kills of birds and small mammals.
It has been suggested that the gray jay might
be a major predator of mammals (Gill 1974). Predation
seems unlikely to be a major source of food for the
jays and may be inflated by a few casual observations.
Gray jays do not have the morphological adaptations to
be efficient predators. Most observations of gray
jays eating meat suggest the animals were previously
killed. Decaying carcasses are probably the greatest
source of meat to the jays, not their own hunting
activities. Decaying meat can be more easily removed
from a carcass than from a fresh killed animal. It is
conceivable that gray jays do, therefore, detect the
scent of rotting meat in nature.
Jays were also tested with fresh ground beef to
determine if they could smell beef that was not
decaying. The results indicated that the jays can
find the fresh beef. However, learning appears to be
involved in finding the fresh beef. The jays showed
improvement over the trials administered, suggesting
that a sensitization period might be required.
One of the birds tested with processed cheese
performed perfectly on all trials, indicating that
this jay could smell cheese. The other two jays
tested did not find the cheese better than random
search. It is possible that birds can become
sensitized to different food items. Genetically based
or early developmental neurologic fixation on scents,
as suggested by Bang and Wenzel (1985), may not occur,
and the jays may have to learn what scents to forage
for. Another explanation for this occurrence is that
the one jay preferred this food and was better
motivated to find it. Food preferences of the jays
were not tested, but searching behavior was consistent
for all food items, and all food items were readily
consumed when found.
The results from the one bird that found cheese
appear to contradict the conclusions of Bunch and
Tomback (1986). These investigators tested for
olfaction with processed cheese and pungent beer
cheese. The gray jays in Bunch and Tomback's study
were not trained to search by olfaction but instead
used spatial memory. Apparently, corvids can be
trained to search using different methods. Buitron
and Nuechterlein (1985) used an area where magpies
naturally foraged and where the investigators fed the
birds. Harriman and Berger (1987) trained their birds
to search for buried substances. It is possible that
training to a paradigm of olfactory searching behavior
is required before birds will use this cue.
A number of cues may be used by birds to
detect food, as demonstrated in the honey bee (Apis
mellifica) (Masuhr and Menzel 1972). Olfaction may be
used by the gray jay when visual cues are not present.
A jay might follow an odor gradient to a food location
when the food is not visible. Once in the area, the
jay may change to a visual searching strategy. Past
work in the lab with gray jays suggest that the jays
may be trained to find food by spatial memory or by
olfaction. From my experience, training jays to
forage by olfaction is not an easy task.
Spatial memory has been shown to exist in
several corvid species (Bossema 1979, Baida 1980,
Tomback 1980, Vander Wall 1982, Kamil and Baida 1985,
Bunch and Tomback 1986). Caution should be used in
the interpretation of spatial memory studies, unless
use of olfaction is controlled. It is possible that
other corvid species are able to use olfaction while
foraging for food items. When demonstrating spatial
memory, the investigator should thoroughly test birds
before ruling out the use of olfaction.
The OB/CH percentage ratio of the gray jay and
other passerines species may not be indicative of
their ability to use olfaction. The role of
olfaction, particularly in behavioral studies, should
not be discounted. This study manifests the
importance of testing animals with ecologically
relevant scents. As suggested by Buitron and
Nuechterlein (1985), the small OB/CH ratio may not
represent a decreased ability to detect scents but
rather a less accurate ability in odor discrimination.
Alternatively, I suggest that small OB/CH ratios
represent good discrimination for a few scents.
Further investigations of olfaction should be pursued
with corvids and other passerine species to determine
the extent of their olfactory abilities. The use of
ecologically relevant scents are critical to these
Baida, R. P. 1980. Recovery of cached seeds by a
captive Nucifraga caryocatactes. Zeitschrift fur
Tierpsvchologie, 52, 331-346.
Bang, B. G. 1971. Functional anatomy of the
olfactory system in 23 orders of birds. Acta
anatomica, 79, Supplement 58, 1-76.
Bang, B. G. and Cobb, S. 1968. The size of the
olfactory bulb in 108 species of birds. Auk, 85,
Bang, B. G. and Wenzel, B. M. 1985. Nasal cavity and
olfactory system. In: Form and Function in Birds
(Ed. by A. S. King and J. McLelland), pp. 195-225.
London: Academic Press.
Bossema, I. 1979. Jays and oaks: an eco-ethological
study of a symbiosis. Behaviour, 70, 1-117.
Buitron, D. and Nuechterlein, G. L. 1985.
Experiments on olfactory detection of food caches
by black-billed magpies. Condor, 87, 92-95.
Bunch, K. G. 1985. Bolus recovery by gray jays: an
experimental analysis. M. A. thesis, University
Bunch, K. G. and Tomback, D. F. 1986. Bolus recovery
by gray jays: an experimental analysis. Animal
Behaviour, 34, 754-762.
Dow, D. D. 1965. The role of saliva in food storage
by the gray jay. Auk, 82, 139-154.
Eckert, K. R. 1987. A documented invasion of gray
jays in Duluth (Minnesota), fall 1986. Loon. 59,
Goodwin, D. 1986. Crows of the World. St.
Edmundsbury Press, England.
Gill, D. 1974. The gray jay as a predator of small
mammals. Canadian Field-Naturalist, 88, 370-371.
Grubb, T. C., Jr. 1974. Olfactory navigation to the
nesting burrow in Leach's petrel Oceanodroma
leucorhoa. Animal Behaviour, 22, 192-202.
Harper, F. 1953. Birds of the Ungava peninsula.
University of Kansas Press, Lawrence, Kansas.
Harriman, A. E. and Berger, R. H. 1986. Olfactory
acuity in the common raven (Corvus corax).
Physiology and Behavior, 36, 257-262.
Hutchison, L. V. and Wenzel, B. M. 1980. Olfactory
guidance in foraging by Procellariiformes.
Condor, 82, 314-319.
Janssen, R. B. 1987. An incursion of gray jays into
southern Minnesota, 1986-87. Loon. 59, 153-154.
Kamil, A. C. and Baida, R. P. 1985. Cache recovery
and spatial memory in Clark's nutcrackers
(Nucifraqa Columbiana). Journal of Experimental
Psychology: Animal Behavior Processes. 11,
Masuhr, T. and Menzel R. 1972. Learning experiments
on the use of side-specific information in the
olfactory and visual system in the honey bee (Apis
mellifica). In: Information Processing in the
Visual Systems of Arthropods (Ed. R. Wehner),
pp. 315-322, Springer-Verlag, Berlin, Germany.
Ouellet, H. 1970. Further observations on the food
and predatory habits of the gray jay. Canadian
Journal of Zoology, 48, 327-330.
Peterson, T. T. 1977. Olfactory sensitivity and
function in the black-billed magpie and in the
pigeon. M. A. thesis, University of Colorado.
Robbins, C. S,, Brunn, B. and Zim, H. S. 1983. Birds
of North America. Golden Press, New York, New
Rutter, R. J. 1969. A contribution to the biology of
the gray jay (Perisoreus canadensis). Canadian
Field-Naturalist 83. 300-317.
Schmidt-Koenig, K. 1987. Bird navigation: has
olfactory orientation solved the problem?
Quarterly Review of Biology, 62, 31-47.
Smith, S. A. and Paselk, R. A. 1986. Olfactory
sensitivity of the turkey vulture (Cathartes aura)
to three carrion-associated odorants. Auk, 103,
Stager, K. E. 1964. The role of olfaction in food
location by the turkey vulture (Cathartes aura).
Los Angeles City Museum Contributions in Science,
Tomback, D. F. 1980. How nutcrackers find their seed
stores. Condor, 82, 10-19.
Vander Wall, S. B. 1982. An experimental analysis of
cache recovery in Clark's nutcrackers. Animal
Behaviour, 30, 84-94.
Waldvogel, J. A. 1987. Olfactory navigation in
homing pigeons: are the current models
atmospherically realistic? Auk, 104, 369-379.
Walker, W. F. 1987. Functional Anatomy of the
Vertebrates: An Evolutionary Perspective.
Saunders College Publishing, New York, New York.