Effects of sex, social status and gonadectomy on countermarking by domestic dogs, Canis familiaris

Abstract

To investigate the social roles of countermarking in dogs, we measured tail base position (TBP, a measure of social status), adjacent-marking and overmarking responses of male, nonoestrous female, gonadectomized male and gonadectomized female dogs to controlled presentations of urine from unfamiliar dogs and social groupmates. We also recorded dog sex, TBP, countermarking and urine investigation at a park. In urine presentations, only males overmarked, intact males (but not gonadectomized males) preferentially overmarked intact female urine, overmarking males had higher TBP than males that did not countermark and urine familiarity did not affect overmarking. In contrast, dogs adjacent-marked only unfamiliar samples, and neither sex nor TBP significantly affected adjacent marking. Gonadectomy did not significantly change the likelihood of countermarking. In dog park observations, males and females marked at and investigated ‘scent posts’ comprised of serial countermarks, often associated with visual landmarks. Males and females were equally likely to countermark and investigate urine and countermarks made up a similarly large portion of countermarking for males and females. Males and females with higher TBP urinated, investigated urine, and countermarked more than same-sex dogs with lower TBP. These studies suggest that although intact males may overmark female urine in part to guard mates as previously hypothesized, both sexes, intact and gonadectomized, likely countermark competitively. Social and sexual patterns also suggest that overmarks and adjacent marks may be distinct signals.

Full text

The effects of sex, gonadectomy and status on investigation patterns of unfamiliar
conspecific urine in domestic dogs, Canis familiaris
Anneke E. Lisberg
a
,
b
,
*
, Charles T. Snowdon
a
,
c
a
Department of Zoology, University of Wisconsin-Madison
b
Department of Biological Sciences, University of Wisconsin-Whitewater
c
Department of Psychology, University of Wisconsin-Madison
article info
Article history:
Received 30 May 2008
Initial acceptance 11 September 2008
Final acceptance 29 December 2008
Published online 14 March 2009
MS. number: A08-00368R
Keywords:
Canidae
Canis familiaris
domestic dog
gonadectomy
olfactory communication
progesterone
scent marking
social hierarchy
testosterone
urine marking
Little is known about social roles of urine marking and mark investigation in interpack communication in
canids beyond mate acquisition, or of the influence of gonadal hormones on interpack urinary
communication. We studied the responses of male and female, intact (nonoestrous) and gonadectomized
Labrador retrievers to urine from unfamiliar dogs of the same four reproductive categories. We compared
two measures of subject social status within subjects’ home groups and measured the effect of subject
status on urine investigation duration. We also compared the responses of subjects to urine from familiar
and unfamiliar dogs. All subjects showed high interest in unfamiliar urine samples, but differed in their
preference for particular unfamiliar urine types. Intact males, intact females and spayed females
investigated urine from intact males and females equally, whereas neutered males investigated urine
from intact males longer than urine from intact females. Urine from intact males and females elicited
longer investigation than did urine from gonadectomized sources. Males showed no difference in
investigation of familiar and unfamiliar urine, whereas females showed a nonsignificant tendency to
investigate urine from unfamiliar dogs longer than urine from groupmates. Low-status subjects inves-
tigated urine longer than high-status subjects. Sex, status and familiarity patterns suggest that males and
females, both intact and gonadectomized, have strong interest in unfamiliar urine and investigate urine
to assess unfamiliar conspecifics in multiple social contexts, including mate and threat assessment. These
results also suggest that gonadal hormones may affect urine investigation patterns both by increasing
sexually motivated urine investigation in males and by creating signals in urine that allow assessment of
potentially risky conspecifics.
Ó2009 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Scent marking serves to reduce and resolve conflict in many
species, aiding both the marker and the mark investigator by
allowing identification of high threat, high status or extremely fit
competitors while avoiding direct confrontations (house mice, Mus
domesticus:Gosling & McKay 1990; Hurst & Rich 1999; lobsters,
Homarus americanus:Karavanich & Atema 1998a, b). Hormones can
play complex roles in this process by contributing to scent mark
composition and influencing marking and investigatory behaviours
(see overview in Wyatt 2003). Androgens in particular can alter
secretion composition (domestic dogs: Dunbar et al. 1980; grey
wolves, Canis lupus:Raymer et al. 1986; meadow voles, Microtus
pennsylvanicus:Ferkin & Johnston 1993) and may correspond to
social status or competitive success, breeding status, and overall
body condition (grey wolves: Asa et al. 1990; cichlid fish, Oreo-
chromis mossambicus:Oliveira et al. 1996; naked mole-rats,
Heterocephalus glaber:Clarke & Faulkes 1998; California mice, Per-
omyscus californicus:Oyegbile & Marler 2005), thereby mediating
chemical signals that honestly reflect these traits. Individuals in
highly social species could assess hormonally mediated scent
marks not only to avoid high-risk interactions, improve competitive
success and choose quality mates, but also to form and maintain
social groups and social hierarchies. However, the relations
between hormones, scent marking and social status in species with
long-term stable social groups and social hierarchies are poorly
understood. Canids provide a unique model for studying the
mechanisms, social roles and evolution of the relations between
hormones, chemical communication and social status in a complex
social system. As a whole, canids provide many examples of
complex social behaviours (Asa 1997), and urine marking may
mediate communication within and between social groups (Anisko
1976 ).
Domestic dogs provide an experimentally flexible model for
studying both the social functions of urine marks and the detailed
inter-relations between gonadal hormones, social status and inter-
and intragroup chemical communication. Studies on dogs may
*Correspondence: A. E. Lisberg, Department of Biological Sciences, University of
Wisconsin-Whitewater, Whitewater, WI 53190, U.S.A.
E-mail address: lisberga@uww.edu (A.E. Lisberg).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/yanbe
0003-3472/$38.00 Ó2009 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.anbehav.2008.12.033
Animal Behaviour 77 (2009) 1147–1154

improve our understanding of these relationships, guide and aid
interpretation of similar studies in wild canids, and have direct
applications for the management and understanding of domestic
companion and free-ranging canids. These studies could especially
complement studies of gonadal hormones and urine marking in
closely related grey wolves (Peters & Mech 1975; Asa et al. 1985,
1990; Mertl-Millhollen et al. 1986; Raymer et al. 1986).
Most studies of urinary behaviours in dogs focus on the relations
between mating, gonadal hormones and urinary behaviours.
Urination rates increase in both sexes in oestrous seasons and with
exogenous gonadal hormone administration (Beach 1974; Dunbar
1978; Pal 2003). Within established groups, urine presentation
studies suggest that urine marking and urine investigation are
important in mating and that males investigate and respond to
urine marks more than females (Doty & Dunbar 1974; Dunbar 1977,
1978; Dunbar et al. 1980; Ranson & Beach 1985). However, urine
from unfamiliar dogs may elicit different responses. Although
aggression and competition occur among groupmates, established
relationships, social hierarchies and appeasement behaviours
decrease intragroup aggression (Scott & Fuller 1965; Schenkel
1967; Beach et al. 1982a). Along with habituation (Dunbar & Car-
michael 1981), established social relationships may contribute to
a decreased need for urine marking and investigation beyond
a mating context among groupmates. Urine from unfamiliar
conspecifics, however, may provide detailed and socially relevant
information about prospective mates, playmates and potential
threats. Until new relationships are established, the threats posed
by unfamiliar conspecifics, as sources of possible aggression and as
competitors, create an incentive for dogs to investigate their urine
to identify or assess unfamiliar individuals.
Gonadal hormones may contribute to the use of urine marks in
unfamiliar conspecific assessment in dogs. In grey wolves, social
status, testosterone and urine-marking rates appear tobe positively
correlated during the breeding season (Peters & Mech 1975; Mertl-
Millhollen et al. 1986; Asa et al. 1990), and exogenous testosterone
administration affects volatile urine components (Raymer et al.
1986). The widespread gonadectomy of domestic dogs provides
a large subject pool in which to test the effects of gonadal hormones
on urine investigation patterns and thereby illuminate the effects of
gonadal hormones on urine signals and signal reception and direct
further study on the contributions of individual gonadal hormones.
In this study, we established the investigatory patterns of dogs
to unfamiliar conspecific urine, assessed whether dogs investigate
unfamiliar conspecific urine outside of an immediate mating
context, and examined the effects of gonadectomy on the responses
to urine from unfamiliar dogs. When possible, we also measured
the investigatory patterns to urine from familiar versus unfamiliar
dogs. We made the following predictions.
Both sexes have an incentive to establish relationships with
unfamiliar conspecifics, so males and females presented with
unfamiliar urine should spend equal time investigating samples.
Dogs are unusual among canids in that female oestrus is asea-
sonal and males maintain consistent testosterone levels and full
breeding capability year-round (Asa 1997). Although we expected
intact males to show a strong interest in unfamiliar female urine
because of sexual interest, we expected that the need to evaluate
and safely establish relationships with unfamiliar males would also
affect males’ investigation patterns. Since male–male aggression
occurs more frequently than male–female aggression in dogs (Scott
& Fuller 1965), and since males are more likely to successfully
outcompete females for limited resources (Scott & Fuller 1965;
Beach 1970; Beach et al. 1982a), unfamiliar males pose a greater
threat than unfamiliar females to male dogs. We therefore expected
that intact males would show strong interest in female urine
because of sexual attraction and all males would display similarly
strong interest in male urine because of the potential threat posed
by unfamiliar males with whom a relationship has yet to be
established. If true, this pattern would differ from that seen in tests
with groupmate urine in which males investigated intact and
spayed female urine much longer than urine from intact males
(Dunbar 1977, 1978; Dunbar et al. 1980).
We expected that gonadectomy of males would reduce males’
sexual interest in female urine, an effect already supported for
females (Dunbar 1978), but would not affect males’ motivation to
investigate urine to assess risk. Therefore, we predicted that neu-
tered males, unlike intact males, would spend more time investi-
gating intact male urine than intact female urine.
Observations on free-ranging dogs suggest that females prefer
familiar dogs as mates (Daniels 1983), and studies on laboratory
beagles showed that spayed females have little interest in urine
from male groupmates until induced into oestrus (Dunbar 1977,
1978). Nonoestrous females may, however, consider mate quality
when assessing unfamiliar males as well as considering the
potential threats that both male and female unfamiliar conspecifics
may pose. Females often lose control of limited resources when in
competition with males (Scott & Fuller 1965; Beach 1970).
However, observations of free-ranging dogs suggest that aggression
from females may be more common than aggression from males
(Pal et al. 1998). We expected spayed and nonoestrous females to
spend equal time investigating urine from unfamiliar intact males
and unfamiliar intact females as a reflection of the potential
benefits and threats posed by unfamiliar dogs of both sexes.
Gonadectomy of urine donors could alter the response to urine
by (1) reducing the ability of subjects to differentiate male and
female urine, (2) reducing the sexual attractiveness of opposite-sex
urine and/or (3) minimizing the chemical cues indicating potential
social threat. If gonadectomy simply confuses the sexual identity of
the urine source, the investigation duration of urine from spayed
females and neutered males should be intermediate to or higher
than the investigation durations of urine from intact males and
females. If gonadal hormones contribute to the sexual attractive-
ness of female urine outside of oestrus, then intact males should
investigate urine from spayed females less than urine from intact
females. If donor gonadectomy renders urine less threatening
either by creating urine similar to that of immature dogs and/or by
interrupting the production of chemical indicators of social status
or body condition, donor gonadectomy will reduce same-sex
investigation by males and same-sex and opposite-sex investiga-
tion by females.
Since habituation and established social relationships may
decrease interest in urine from groupmates, and studies of free-
ranging dogs suggest that intergroup aggression is more common
than intragroup aggression (Pal et al. 1998), we predicted that dogs
would spend more time investigating urine from unfamiliar indi-
viduals than from groupmates.
We expected individuals of lower status to spend more time
investigating conspecific urine than same-sex high-status individ-
uals since low-status individuals may be more likely to be out-
competed for resources by unfamiliar dogs than high-status
individuals, and may therefore be more likely to view unfamiliar
conspecifics as a potential threat.
EFFECTS OF FROZEN STORAGE ON RESPONSE TO URINE
To determine the viability of using previously frozen and
thawed urine samples in the subsequent studies, we first compared
the responses of dogs to fresh and frozen, thawed urine samples.
Previous studies have used frozen samples or overnight collection
methods, but the effects of freezing urine on subsequent responses
have not been tested relative to freshly collected urine.
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–1154114 8

Methods
Eight dogs (four spayed females, four neutered males) served as
subjects. All were at least 1 year of age at the time of the study. We
used several breeds, including mixed breeds, but excluded
brachycephalic dogs to avoid any potential effects of genetically
reduced muzzles on olfaction.
Eight additional dogs (four intact males, four intact females)
served as urine donors. All donors were at least 1 year of age at the
time of the study. Females showing signs of oestrus or pro-oestrus
were excluded. Donors comprised multiple breeds; however, we
excluded Dalmatians (Safra et al. 2006) and bulldog breeds (Bartges
et al. 1994) because of the high incidence of uroliths reported in
these breeds. We collected urine samples noninvasively using
a sterile nonreactive plastic urine cup held in the urine stream. We
collected urine between 1000 and 1700 hours, but excluded the
first urination of the morning. We presented a portion of the fresh
urine samples to subjects and froze the remaining urine in the
collection cup within 3 h of collection. These samples remained
frozen at 25

C for 7–28 days until thawed at room temperature
for use in frozen-sample presentation.
We presented each subject with two urine samples from an
intact male donor and two samples from an intact female donor,
with urine from each donor presented once while fresh and again
after freezing. Urine samples were presented 2.85 m apart, and
each sample was presented just once before freezing and once after
freezing to each subject. We presented urine from each donor to
two subjects, one male and one female. For each presentation, we
spread urine with a cotton swab over a 12 3 cm area on a 29 cm
long raw wooden stake placed upright in the ground to amplify the
signal and ensure that subjects could locate the source. Subjects’
owners, blind to the samples, walked each subject up to the stake
on-leash to allow investigation of each sample. Each session ended
when the subject walked away from the stake following investi-
gation. We videorecorded each investigation and measured inves-
tigation duration based on the time that the subject held its nose
within approximately 10 cm of the stake. When possible (4 of 8
subjects), we presented both male and female samples simulta-
neously to each subject.
Results
We performed a sign test using each of the 16 fresh-frozen pairs
(one pair from an intact male and one pair from an intact female for
each subject) and found no significant difference in investigation
duration of fresh and frozen samples (N¼16, P¼0.118). We also
found a strong correlation between investigation durations of fresh
and frozen presentations of each sample (Pearson’s correlation:
r
14
¼0.520, P¼0.039). These results support the efficacy of frozen
samples as stimuli.
EFFECTS OF SEX, GONADECTOMY, SOCIAL STATUS AND
FAMILIARITY ON URINE INVESTIGATION
We observed the investigation of urine samples from unfamiliar
intact males, intact females, neutered males and spayed females by
Labrador retrievers of the same-sex and gonadal status categories
and recorded two measures of social status for each subject.
Methods
We studied 40 privately owned Labrador retrievers as subjects:
10 intact males, 10 intact females, 10 spayed females and 10 neu-
tered males. All dogs were at least 1 year of age at the time of the
study. Spayed and neutered dogs were gonadectomized at 6
months or older (7 males were neutered between 6 months and 1
year, 3 males were neutered after 1 year, 5 females were spayed
between 6 months and 1 year, and 5 females were spayed after 1
year.) We excluded oestrous and pro-oestrous females both as
subjects and as urine donors to isolate the investigatory patterns of
dogs outside of the breeding context, basing nonoestrus state on
both visual inspection and owner/breeder interview. We did not
differentiate anoestrous and metoestrous females since females in
both states are neither sexually receptive nor proceptive (Beach
et al. 1982b) and are likely to have similar interests in investigating
urine for mate appraisal, threat appraisal and other social purposes.
To obtain data on the social status of subjects in their home groups
and to ensure a minimal level of previous social experience with
conspecifics and their urine, all subjects lived with at least one
other dog. Group sizes ranged from 2 to 14 (mean SD group
size ¼5.25 3.00; see Table 1).
We collected urine from intact males, intact females, neutered
males and spayed females that were unfamiliar to the subjects. We
collected urine from 48 dogs, including Labrador retrievers, grey-
hounds and border collies. When possible, we also collected urine
from the groupmates of the subjects. We did not obtain adequate
urine samples from all groupmates on test days, and therefore
included urine from 75 groupmates (21 intact males, 18 intact
females, 16 neutered males, 20 spayed females) of only 31 subjects
(9 intact males, 8 intact females, 7 neutered males, 7 spayed
females).
Urine collection occurred as in experiment 1, with unfamiliar
samples kept frozen at 25

C for up to 28 days before thawing at
room temperature for use. We presented urine from groupmates
within 1 h of collection.
Urine course
We applied urine or tap water samples to wooden stakes, and
placed six stakes (four stakes with urine samples from unfamiliar
dogs representing each of the four donor categories plus two
control stakes with water) 2.85 m apart in a row, aligned perpen-
dicular to the prevailing wind, outside the home of the subject. We
randomized the stake presentation order with the constraint that
each sample type occurred in each position at least once over the
course of the study. Following exposure to the unfamiliar and
control samples, we presented additional wooden stakes with
urine samples from groupmates using the same presentation
methods.
Either the subjects’ owners (for 25 subjects) or the primary
author (for 15 subjects) served as a handler and led each subject
individually through the urine course on-leash, with other dogs
kenneled or housed out of view. Handlers stopped within 90 cm of
each stake in turn and moved on to the next stake after 30 s when
cued by a second observer, unless the subject was actively inves-
tigating the stake or urinating at the end of 30 s. Unless the subject
attempted to move on to the next stake before the 30 s, handlers
maintained a loose leash and did not speak, gesture, or in any way
guide the subject. Other than maintaining basic proximity to each
stake, handlers made no attempt to direct the subjects’ attention to
Table 1
Group composition of subjects by sex and gonadal status
Subject category Subject housed with
Intact male
groupmates
Intact female
groupmates
Neutered male
groupmates
Spayed female
groupmates
Intact males 8 7 2 5
Intact females 10 9 4 5
Neutered males 3 4 7 5
Spayed females 6 5 3 9
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–1154 114 9

the sample. Handlers were blind to stake order and sample content
for all but five subjects (2 intact females, 2 spayed females and 1
intact male).
The primary author or a second observer recorded and video-
taped the responses in the urine course so responses could be
determined from viewing tapes. The observer stood at least 6 m
from the nearest stake, timed the subject at each stake, and gave
brief verbal cues to notify the handler to move to the next stake.
From the videotapes, we recorded the duration of investiga-
tion of each stake, defined as the subject holding its nose within
approximately 10 cm of the stake. Although exact measurement
of distance from the stake was not possible, subjects clearly and
distinctly withdrew from each stake at investigation termina-
tion, serving as a secondary indicator. When subjects investi-
gated the sample multiple times, we summed durations of all
investigations unless an investigation occurred after the subject
urinated on the sample. A second observer blind to sample order
independently scored 20% of subjects, and we compared our
scores of investigation duration for each stake to ensure accu-
racy (Pearson’s correlation: r
46
¼0.945, P<0.001). Rank order
scores for stake investigation duration were the same for all
subjects.
Status tests
We determined social status for each subject using two inde-
pendent measures: (1) mean tail base position through the urine
course and (2) a possession score, where high status was defined as
an individual having priority of access to a desired and limited
resource (a novel, food-filled toy). We measured the ability of the
subjects to obtain and maintain possession of the toy when
groupmates were present, unless groupmates were excluded by the
owner (N¼2.) After handlers led the subjects through the urine
course, we videotaped the entire home group of dogs off-leash as
we allowed them to smell the toy. We then threw the toy three
times. For each throw, any dog that obtained and successfully
maintained possession of the toy for 30 s (or until the dog inde-
pendently returned the toy to the owner) received a score of 1. Each
subject earned a possession score between 0 (no possessions) and 3
(possessed toy in all three throws). Some dog owners preferred not
to include some or all of their dogs in this test, so we tested 33
subjects (7 intact males, 8 neutered males, 8 intact females and 10
spayed females) in this way.
This test may not, however, accurately measure status for all
individuals, since variation in aggression level, motivation for the
object of the competition, sex composition of group and ability to
obtain first possession (which, for this study, could be influenced by
mobility or proximity to thrown toy) can influence possession tests
(Scott & Fuller 1965) and mask true status. To address these
concerns, we scored tail carriage for each subject on approach to
each stake in the urine course as low (1), medium (2), or high (3),
and averaged the scores for each subject across the urine course as
a second indicator of status.
Although tail postures in wolves are clearly linked to status
(Harrington & Asa 2003), and these same tail postures occur in dogs
(Scott & Fuller 1965), the relation between priority of access to
limited, desired resources and tail postures has not been tested in
dogs. We analysed agreement between mean tail base position and
possession score for each subject. A second researcher blind to
sample order also independently scored tail base positions for 15%
of the subjects and compared the results to our scores to ensure
objectivity and accuracy of scoring (Pearson’s correlation:
r
34
¼0.890, P<0.001). For both status tests we made no assump-
tion about absolute position (such as alpha or beta), and in some
cases, multiple individuals in a home group received the same
status score using each measure.
Analysis
To determine whether the investigation durations in response
to the urine samples represented a response to the urine itself
rather than a response to a novel wooden stake, we compared the
mean investigation durations of the four urine samples and the two
water samples for each subject using a Wilcoxon signed-ranks test.
Presentation order could affect the duration of investigation
because of novelty effects (the first stake might elicit a longer
duration than the last) or because of learned association (later
stakes might elicit longer durations as subjects learn to associate
urine samples with the stakes). To assess the effect of stake order on
investigation duration, we performed a Friedman ANOVA on the
duration of investigation of each stake position (1–6) for all
subjects.
To identify systematic differences in overall investigation
duration across all four subject categories, we used a Kruskall–
Wallis test to compare the longest investigation duration of a single
unfamiliar urine stake by intact males, intact females, neutered
males and spayed females. We focused on the longest duration of
investigation of a single stake by each subject rather than mean
investigation duration of all four urine stakes per subject to elim-
inate the potential effect of selectivity of various stimuli.
Having shown equivalent responsiveness across subject condi-
tions, we then looked for differences in responsiveness to odour
types within subject conditions. We used Friedman’s ANOVA to test
the effect of reproductive category of urine donor on the duration of
investigation for each subject group. For those subject categories
that displayed a significant main effect of urine donor category, we
performed additional Wilcoxon signed-ranks tests to compare the
percentage of total investigation duration spent on urine from
intact male and intact female donors. Since gonadal hormones
could affect urine composition differently in males and females, to
test the effects of donor gonadectomy on response to urine, we
used separate Wilcoxon signed-ranks tests to compare the
percentage of investigation duration of urine from intact males and
neutered males, and to compare the percentage of investigation
duration of urine from intact and spayed females. Because of the
high variation in investigation duration between subjects, we
scored the response to each unfamiliar urine sample as
a percentage of the subject’s total time spent investigating the four
unfamiliar urine stakes to focus on relative interest of subjects in
each reproductive category. We then arcsine-transformed the
percentage scores to approximate normality for each Wilcoxon test.
We used a Wilcoxon signed-ranks test to compare duration of
investigation of urine from groupmates and from unfamiliar dogs of
the same reproductive category as the groupmates. When multiple
groupmates belonged to the same reproductive category, we used
the mean duration of all groupmates in the same category. We
evaluated males and females separately, since sexual attraction,
aggression and potential threat from groupmates and unfamiliar
conspecifics might all differ between the sexes and could affect
investigation patterns.
To test the validity of using tail base position as a measure of
social status, we used a Kruskall–Wallis test followed by Mann–
Whitney Utests to determine whether dogs that did not obtain the
toy in the possession test with groupmates displayed a lower mean
tail base position in the urine course than dogs that always
obtained the toy. To determine whether dogs of lower status
investigated unfamiliar urine longer than did dogs of higher status,
we performed a linear regression of the longest investigation
duration on tail base position recorded in the urine course. We used
the longest investigation duration recorded for any unfamiliar
sample in the urine course rather than mean investigation duration
to differentiate interest level from selectivity.
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–1154115 0

Ethical note
Both experiments were conducted under the approval of the
University of Wisconsin-Madison College of Letters and Science
Institutional Animal Care and Use Committee (protocol code
L00352).
Results
Subjects investigated the four wooden stakes with urine from
unfamiliar dogs significantly more than the two stakes treated with
water (mean SE investigation duration: unfamiliar urine:
5.12 0.68 s; water: 0.50 0.11 s; Wilcoxon signed-ranks test:
Z¼5.511, N¼40,P<0.001). Stake order did not affect the duration
of sample investigation (Friedman ANOVA:
c
5
2
¼6.039, P¼0.302).
Subjects across the four reproductive categories showed no
difference in longest investigation duration of unfamiliar urine
samples (Kruskall–Wallis test:
c
3
2
¼0.689, P¼0.876). The repro-
ductive category of the donor affected duration of investigation
only for neutered male subjects (Friedman’s analysis:
c
3
2
¼8.548,
P¼0.036) and intact female subjects (
c
3
2
¼8.443, P¼0.038).
Neutered males spent a higher percentage of total investigation
time with urine from intact males than with urine from intact
females (Wilcoxon signed-ranks test: Z¼1.988, N¼10 ,
P¼0.047; Fig. 1), whereas intact males, intact females and spayed
females investigated urine from intact males and females equally
(Fig. 1). Although intact females investigated urine from intact
males and females equally, they responded differently to urine
from intact and gonadectomized sources.
Overall, subjects investigated urine from intact males longer
than urine from neutered males (Wilcoxon signed-ranks test:
Z¼2.844, N¼40, P¼0.004), and they investigated urine from
intact females longer than urine from spayed females (Z¼2.313,
N¼40, P¼0.021). The difference in investigation time of urine
from intact and neutered males was significant for intact and
neutered male subjects considered collectively (Z¼2.073, N¼20,
P¼0.038), neutered male subjects alone (Z¼1.988, N¼10,
P¼0.047; Fig. 1) and intact female subjects alone (Z¼2.293,
N¼10, P¼0.022; Fig. 1). The difference in investigation time of
urine from intact and spayed females was not significant for intact
and spayed female subjects considered collectively (Z¼1.737,
N¼20, P¼0.082), or for intact male subjects alone (Z¼1.7 18 ,
N¼10, P¼0.086; Fig. 1). Since neutered males showed low interest
in intact female urine, the response of neutered males to intact and.
spayed female urine could not be assessed.
Females tended to investigate urine from unfamiliar conspe-
cifics longer than they did urine from groupmates of the same
reproductive categories, but the difference was not significant
(Z¼1.7 64, N¼20, P¼0.078; Fig. 2). Males showed no difference
between familiar and unfamiliar samples.
There was strong agreement between mean tail base position in
the urine course and possession score in the status test within the
subjects’ home group (Kruskall–Wallis test:
c
3
2
¼12.390, P¼0.006;
Fig. 3). The mean difference in tail base position between subjects
that received a score of 0 (0 of 3 possessions) and a score of 3 (3 of 3
possessions) in the status test was significant (Mann–Whitney U
test: Z¼3.411, N¼22, P<0.001). Dogs of lower status investigated
unfamiliar urine longer than did dogs of higher status, using mean
tail base position as a measure of status (linear regression:
R¼0.376, F
1,138
¼6.275, N¼40, P¼0.017; Fig. 4).
DISCUSSION
The brief investigation duration of control stakes relative to
those treated with urine indicates that investigation duration was
a viable measure of interest in samples presented on stakes.
Although duration was brief, the investigation of control stakes
may reflect general interest in vertical objects, which could
improve localizability of urine marks. However, novelty and
learned association of urine with the stakes also may have
contributed to the brief response to the control stakes.
Dunbar (1978) presented beagles with urine from groupmates,
and found that males investigated urine much longer than females
did: 11 times longer with oestrous female urine, 7 times longer
with spayed female urine, and 2 times longer with male urine. In
contrast to the strong sexual dichotomy reported by Dunbar (1978)
and Ranson & Beach (1985), we found that males and females, both
Intact male urine
Neutered male urine
Intact female urine
Spayed female urine
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Intact males
Neutered males
Intact females
Spayed females
Sub
j
ect cate
g
or
y
Proportion of investigation
a
b
a
b
b
Figure 1. Mean þSE proportion of time that subjects spent investigating urine from
unfamiliar dogs in each reproductive category. Although proportions of total investi-
gation time are shown here, statistical analysis was conducted using arcsine-trans-
formed percentages of total investigation time to approximate normality. Different
superscript letters above bars denote a significant difference in investigation time
(P<0.05).
Familiar
Unfamiliar
7
6
5
4
3
2
1
0
Duration of investigation (s)
Males Females
Sub
j
ect cate
g
or
y
Figure 2. Mean þSE time that male and female subjects spent investigating urine
from familiar and unfamiliar conspecifics.
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–1154 1151

intact and gonadectomized, showed an equally strong interest in
the urine of unfamiliar dogs. Although it is possible that there are
effects of breed or rearing, the most likely cause of the different
results between these studies is the familiarity of the urine sources
used. Dunbar (1978) and Ranson & Beach (1985) presented urine
from dogs housed with the subjects, whereas we tested responses
to urine from familiar and unfamiliar dogs. The nonsignificant trend
by females in this study towards investigating unfamiliar urine
longer than urine from groupmates and the longer investigations of
unfamiliar urine (equal to males’ investigations) further support
the possibility that females respond differently to urine from
familiar and unfamiliar conspecifics, and that while males may be
more active urine investigators of intragroup urine than females,
both sexes have equal interest in unfamiliar urine.
Neutered males investigated urine from intact males more than
urine from any other unfamiliar dogs, whereas females (both intact
and spayed) investigated urine from intact males and females
equally. Although individuals varied substantially, these patterns
(along with patterns of social status and urine investigation dis-
cussed below) support the idea that dogs investigate urine in part
to assess unfamiliar and potentially high-threat individuals, since
males are most likely to lose resources to other males, whereas
females may lose resources to both males and females. Investi-
gating unfamiliar urine might simply alert dogs to the presence of
potentially high-threat individuals, or it may provide more detailed
information about urine source identity, social status, health, or
body condition.
Intact males investigated urine from intact males and females
equally. This response differs from the strong preference for female
urine (even urine from spayed females) over urine from intact
males seen by Dunbar (1978) when males were presented with
urine from groupmates. The high interest that intact males showed
to intact male and female urine in our study suggests that although
intact males investigated familiar and unfamiliar female urine to
assess potential mates, males also investigated urine from unfa-
miliar males for other purposes. Males may investigate same-sex
conspecific urine to assess potential competitors, sources of
potential aggression, and to otherwise aid in the safe establishment
of new relationships.
The lower investigation duration of intact female urine dis-
played by neutered males relative to intact males suggests that
gonadal hormones contribute to male interest in urine from dogs of
the opposite sex, a result that was established for females by
Dunbar (1978). Gonadal hormones do not, however, appear to
affect the motivation of dogs of either sex to investigate urine from
unfamiliar conspecifics for other purposes, including risk assess-
ment, since intact and gonadectomized subjects showed equal
interest in unfamiliar urine and displayed strong interest in urine
from intact males and females, and since low-status subjects across
all four reproductive categories investigated unfamiliar urine
longer than did high-status subjects.
The reduced interest in intact female urine displayed by neu-
tered males relative to intact males may be influenced by a lack of
sexual experience in neutered males. Although Doty & Dunbar
(1974) found that sexual experience is necessary for males to
differentiate between urine from spayed females hormonally
induced into false oestrus and untreated spayed females, we found
no evidence that sexual experience induced intact males’ high
interest in intact female urine. Of the eight intact male subjects
whose investigation time of intact male and intact female urine
differed by more than 5%, three investigated intact male urine more
than intact female urine (i.e. similar to responses of neutered
males). Two of these males had mating experience but one did not.
Of the five intact males that investigated urine of intact females
more than urine of intact males, three had mating experience and
two did not.
Dunbar (1978) found that intactmales showed a greater interest
in urine from oestrous females than urine from spayed females. In
this study, urine from gonadectomized dogs generally elicited
weaker responses than urine from intact dogs even though estrous
females were not included. Although we found no significant
differences in investigation of urine from intact and gonadecto-
mized donors when we considered each subject category sepa-
rately, our finding that urine from gonadectomized dogs (both male
and female) elicited shorter investigations overall than urine from
intact dogs suggests that gonadal hormones may be necessary for
dogs to identify the urine source as having strong social relevance,
including posing a social threat. The role of gonadal hormones in
the production of ‘high-threat’ urine may simply indicate that dogs
are unable to identify the urine of gonadectomized dogs as coming
from adult dogs. Since adult free-ranging dogs show more
3
2.5
2
1.5
1
012
Dominance score with home
p
ack
Mean tail base position (1–3)
a
b
3
Figure 3. Tail base position of subjects of different rank when approaching samples in
the urine course. Each box shows the median (horizontal line), quartiles (boxes/bars)
and extreme values within a category (whiskers). Open circles denote outliners. Rank
of subjects was determined by each subject’s ability to obtain possession of a toy
during competitive interactions with groupmates. Different superscript letters above
bars denote a significant difference in tail base position (P<0.001).
30
20
10
0
1 1.5 2 2.5 3
Mean tail base
p
osition (1–3)
Longest investigation (s)
Figure 4. Longest investigation time of a single urine sample from an unfamiliar
individual by each subject relative to each subject’s mean tail base position during the
urine course (R¼0.376, P¼0.017).
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–11541152

aggression than juveniles (Pal et al. 1998), the urine of apparently
juvenile dogs, while still of some interest and social significance,
may be viewed as a lower-threat signal. However, the great varia-
tion that subjects displayed in response to all urine types combined
with the strong olfactory capabilities of dogs and evidence of
interplay between status, androgens and chemical communication
in grey wolves (Asa et al. 1990) make finer distinctions in urine
investigation extremely likely. The necessary role of gonadal
hormones in creating socially relevant urine could also indicate the
presence of an androgen-mediated indicator of status in urine such
that dogs perceive the urine of gonadectomized dogs as coming
from low-status dogs, or alternatively, that there is an androgen-
mediated indicator of body condition in urine. Further study is
needed to determine whether dogs respond differently to urine
from high- and low-status individuals.
Similar responses to urine have been observed in mice. Major
urinary proteins (MUPs) isolated from the urine of gonadally intact
males appear to serve as strong male aggression-inducing phero-
mones, whereas whole urine from castrated males fails to induce
an aggressive response (Chamero et al. 2007). Male MUPs are also
used by female mice, which integrate MUP information with indi-
cators of status to recognize male urine markers and to inform mate
choice (Nodari et al. 2008). Similarly, dogs may use urinary cues
mediated by gonadal hormones along with other urinary cues to
avoid risky interactions and/or to aid in the safe establishment of
new, positive social relationships.
The different responses to urine from intact and gonadecto-
mized dogs have several implications for interpreting urine
presentation studies as well as understanding the chemical
communication of companion dogs. Dogs do differentiate between
urine from intact, nonoestrous females and spayed females. Since
we did not separate anoestrous from metoestrous urine donors in
this study, dogs may not distinguish between urine from anoes-
trous, intact females and urine from spayed females. However, until
this specific comparison is tested, urine from spayed females
should not be assumed to approximate anoestrous female urine.
The reduced interest in urine from gonadectomized male and
female dogs may indicate that these dogs are no longerperceived as
threats or as potential mates based on their urine, or it may be that
their urine is less recognizable as conspecific urine. In either case,
the effects of gonadectomy on urine composition may alter social
introductions and interactions between unfamiliar dogs.
The greater response by females to urine from unfamiliar dogs
fits predictions, and supports the use of urine investigation in part
to assess potentially high-threat conspecifics. However, males did
not respond as predicted, showing no difference in investigation of
familiar and unfamiliar urine. Males may need to monitor urine of
opposite-sex groupmates more intently than females to anticipate
oestrus. Sexual differences in intragroup social dynamics may also
necessitate more consistent intrasexual urine monitoring among
male groupmates than among female groupmates, for example, if
males more actively maintain status hierarchies.
Dogs of lower social status investigated urine from unfamiliar
dogs longer than did dogs of higher status. The different responses
by dogs of high and low status suggest that individual dogs most at
risk of losing resources to conspecifics may spend the longest time
investigating urine from unfamiliar dogs and further supports an
overall pattern of urine investigation as potential risk assessment.
The strong correlation shown between tail base position
maintained through the urine course and successful number of toy
possessions obtained in competition within the home group
supports the validity of using both methods to measure social
status in dogs. This is the first empirical evidence for the relation-
ship between high tail posture and priority access to a limited,
desired resource in domestic dogs. Although dogs use the same
range of tail postures that wolves use in status displays (Harrington
& Asa 2003), dogs differ from wolves in having less rigid social
hierarchies than wolves. Our results, however, indicate that despite
these differences, status signals such as tail base position corre-
spond to a real difference in ability to access limited resources, and
that these signals persist even when groupmates are absent.
Measuring tail base position as an indicator of social status offers
several improvements over object possession tests. Tail base
measures not only provide continuous data that may illuminate
more subtle status differences, but also avoid aggressive encounters
among groupmates, and are not confounded by differing prefer-
ence strengths by subjects for contested objects.
By expanding the analysis of social, sexual and hormonal
influences on urine investigatory patterns to include unfamiliar
urine sources and a full suite of intact, gonadectomized, male and
female subjects and urine sources, we have developed a broader
picture of the role of urine in canine chemical communication.
Future investigation may illuminate further the hormonal and
social factors that affect urinary communication cues and the role
of urine marking and investigation on social introductions between
unfamiliar dogs. Such inquiry will not only aid our understanding
and management of companion and free-ranging dogs, but may
also shed light on possible mechanisms of the establishment and
maintenance of social hierarchies and conflict resolution in other
highly social species.
Acknowledgments
Patricia B. McConnell provided valuable input into study design
as well as canine expertise, and her guidance and support led to the
undertaking of this project. Jeffrey R. Baylis contributed to early
study design, and Stephen C. Gammie, Catherine A. Marler, Lauren
V. Riters, Antony Stretton and Ian Dunbar provided useful feedback.
We thank dog owners who volunteered their time and their dogs as
subjects and urine donors and, finally, Jenna Buley and Andrea Holt
for their assistance in data collection and analysis. Supported in
part by Emlen and Bunde funds of the Department of Zoology,
University of Wisconsin-Madison.
References
Anisko, J. 1976. Communication by chemical signals in Canidae. In: Mammalian
Olfaction, Reproductive Processes, and Behavior (Ed. by R. Doty), pp. 283–293.
New York: Academic Press.
Asa, C. S. 1997. Hormonal and experiential factors in the expression of social and
parental behavior in canids. In: Cooperative Breeding in Mammals (Ed. by
N. Solomon & J. French), pp. 129–149. Cambridge: Cambridge University Press.
Asa, C. S., Mech, L. D. & Seal, U. S. 1985. The use of urine, faeces, and anal gland
secretions in scent-marking by a captive wolf (Canis lupus) pack. Animal
Behaviour,33, 750–760, doi:10.1016/S0003-3472(85)80043-9.
Asa, C. S., Mech, L. D., Seal, U. S. & Plotka, E. D. 1990. The influence of social and
endocrine factors on urine-marking by captive wolves (Canis lupus). Hormones
and Behavior,24, 497–509, doi:10.1016/0018-506X(90)90 038-Y.
Bartges, J. W., Osborne, C. A., Lulich, J. P., Unger, L. K., Koehler, L. A., Bird, K. A.,
Clinton, C. W. & Davenport, M. P. 1994. Prevalence of cystine and urate uroliths
in bulldogs and urate uroliths in Dalmations. Journal of the American Veterinary
Medical Association,204, 1914–1918.
Beach, F. A. 1970. Coital behaviour in dogs. VIII. Social affinity, dominance and
sexual preference in the bitch. Behaviour,36, 131–147.
Beach, F. A. 1974. Effects of gonadal hormones on urinary behavior in dogs. Phys-
iology and Behavior,12, 1005–1013, doi:10.1016/0031-9384(74)90148-6.
Beach, F. A., Buehler, M. G. & Dunbar, I. F. 1982a. Competitive behavior in male,
female, and pseudohermaphroditic female dogs. Journal of Comparative and
Physiological Psychology,96, 855–874.
Beach, F. A., Dunbar, I. F. & Buehler, M. G. 1982b. Sexual characteristics of female
dogs during successive phases of the ovarian cycle. Hormones and Behavior,16,
414–442, doi:10.1016/0018-506X(82)90050-2.
Chamero, P., Marton, T. F., Logan, D. W., Flanagan, K., Cruz, J. R., Saghatelian, A.,
Cravatt, B. F. & Stowers, L. 2007. Identification of protein pheromones that
promote aggressive behaviour. Nature,450, 899–903, doi:10.1038/nature05997.
Clarke, F. M. & Faulkes, C. G. 1998. Hormonal and behavioural correlates of male
dominance and reproductive status in captive colonies of the naked mole-rat,
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–1154 1153

Heterocephalus glaber.Proceedings of the Royal Society of London, Series B,265,
1391–1399.
Daniels, T. J. 1983. The social organization of free-ranging urban dogs II: estrous
groups and the mating system. Applied Animal Ethology,10, 365–373,
doi:10.1016/0304-3762(83)90185-2.
Doty, R. & Dunbar, I. 1974. Attraction of beagles to conspecific urine, vaginal and
anal sac secretion odors. Physiology and Behavior,12, 825–833, doi:10.1016/
0031-9384(74)90020-1.
Dunbar, I. 1977. Olfactory preferences in dogs: the response of male and female
beagles to conspecific odors. Behavioral Biology,20, 471–481, doi:10.1016/
S0091-6773(77)91079-3.
Dunbar, I. 1978. Olfactory preferences in dogs: the response of male and female
beagles to conspecific urine. Biology of Behaviour,3, 273–286.
Dunbar, I. & Carmichael, M. 1981. The response of male dogs to urine from other
males. Behavioral and Neural Biology,31, 465–470, doi:10.1016/S0163-
1047(81)91546-6.
Dunbar, I., Buehlar, M. & Beach, F. 1980. Developmental and activational effects of
sex hormones on the attractiveness of dog urine. Physiology and Behavior,24,
201–204, doi:10.1016/0031-9384(80)90074-8.
Ferkin, M. H. & Johnston, R. E. 1993. Roles of gonadal hormones in control of 5
sexually attractive odors of meadow voles (Microtus pennsylvanicus). Hormones
and Behavior,27, 523–538.
Gosling, L. M. & McKay, H. V. 1990. Competitor assessment by scent matching: an
experimental test. Behavioral Ecology and Sociobiology,26, 415–420.
Harrington, F. & Asa, C. 2003. Wolf communication. In: Wolves: Behavior, Ecology,
and Conservation (Ed. by L. Mech & L. Boitani), pp. 66–103. Chicago: University
of Chicago Press.
Hurst, J. L. & Rich, T. J.1999. Scent marks as competitive signals of mate quality. In:
Advances in Chemical Signals in Vertebrates (Ed. by R. E. Johnston, D. Mu
¨ller-
Schwarze & P. W. Sorensen), pp. 309–314. New York: Kluwer Academic/Plenum.
Karavanich, C. & Atema, J. 1998a. Individual recognition and memory in lobster
dominance. Animal Behaviour,56, 1553–1560.
Karavanich, C. & Atema,J. 1998b. Olfactory recognition of urine signals in dominance
fights between male lobster, Homarus americanus.Behaviour,135, 719–730.
Mertl-Millhollen, A., Goodmann, P. & Klinghammer, E. 1986. Wolf scent marking
with raised-leg urination. Zoo Biology,5, 7–20.
Nodari, F., Hsu, F., Fu, X., Holekamp, T. F., Kao, L., Turk, J. & Holy, T. E. 2008.
Sulfated steroids a s natural ligands of mouse pheromone-sensing neurons.
Journal of Neuros cience,28, 6407– 6418, doi:10.1523/JNEUROSCI.1425-
08.2008.
Oliveira, R. F., Almada, V. C. & Canario, A. V. M. 1996. Social modulation of sex
steroid concentrations in the urine of male cichlid fish Oreochromis mossam-
bicus.Hormones and Behavior,30, 2–12.
Oyegbile, T. O. & Marler, C. A. 2005. Winning fights elevates testosterone levels in
California mice and enhances future ability to win fights. Hormones and
Behavior,48, 259–267.
Pal, S. 2003. Urine marking by free-ranging dogs (Canis familiaris) in relation to sex,
season, place and posture. Applied Animal Behaviour Science,80, 45–59,
doi:10.1016/S0168-1591(02)00178-8.
Pal, S., Ghosh, B. & Roy, S. 1998. Agonistic behaviour of free-ranging dogs (Canis
familiaris) in relation to season, sex and age. Applied Animal Behaviour Science,
59, 331–348, doi:10.1016/S0168-1591(98)00108-7.
Peters, R. & Mech, L. 1975. Scent-marking in wolves. American Scientist,63,
628–637.
Ranson, E. & Beach, F.1985. Effects of testosterone on ontogeny of urinary behavior
in male and female dogs. Hormones and Behavior,19, 36–51.
Raymer, J., Wiesler, D., Novotny, M., Asa, C., Seal, U. & Mech, L. 1986. Chemical
scent constituents in urine of wolf (Canis lupus) and their dependence on
reproductive hormones. Journal of Chemical Ecology,12, 297–314.
Safra, N., Schaible, R. & Bannasch, D. 2006. Linkage analysis with an interbreed
backrcross maps Dalmatian hyperuricosuria to CFA03. Mammalian Genome,17,
340–345.
Schenkel, R. 1967. Submission: its features and function in the wolf and dog.
American Zoologist,7, 319–329.
Scott, J. & Fuller, J. 1965. Genetics and the Social Behavior of the Dog. Chicago:
University of Chicago Press.
Wyatt, T. 2003. Pheromones and Animal Behaviour: Communication by Smell and
Taste. Cambridge: Cambridge University Press.
A.E. Lisberg, C.T. Snowdon / Animal Behaviour 77 (2009) 1147–11541154