See discussions, stats, and author profiles for this publication at:

https://www.researchgate.net/publication/308039627

Reproductive Endocrinology and Musth Indicators in a

Captive Asian Elephant (Elephas maximus)

Article  in  Psychological Reports · September 2016

DOI: 10.1177/0033294116667092

CITATION READS

1 64

3 authors, including:

Charles I Abramson
Oklahoma State University - Stillwater
226 PUBLICATIONS   1,906 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

gene ontology analysis of the response to acute ethanol stress by honey bees, View project

Vocational Education and Training Project View project

All content following this page was uploaded by Charles I Abramson on 26 October 2017.

The user has requested enhancement of the downloaded file.

Article

Reproductive Psychological Reports

0(0) 1–22
Endocrinology and ! The Author(s) 2016
Reprints and permissions:
Musth Indicators in a sagepub.com/journalsPermissions.nav
DOI: 10.1177/0033294116667092
Captive Asian Elephant prx.sagepub.com

(Elephas maximus)
Connie Duer and Tom Tomasi
Department of Biology, Missouri State University, Springfield, MO, USA

Charles I. Abramson
Department of Psychology, Oklahoma State University, Stillwater, OK, USA

Abstract
Even in the best situations, the artificial social constructs of captivity alter natural
elephant behavior and unfortunately create distress. Asian elephants are powerful
and intelligent animals that require consideration for their well-being and prudent
management. The males present particular difficulties due to a temporary state of
heightened aggressive behavior unique to male elephants called ‘‘musth.’’ When he is
in this state, the danger the elephant poses to other animals and the people around
him is considerable. In addition to antagonistic behavior, musth is also characterized
by temporal gland secretion and urine dribbling. In previous studies, musth has been
attributed to elevated testosterone levels. This study attempted to enhance the
knowledge base concerning these phenomena by examining hormone concentrations
(n ¼ 357) in Onyx, a male Asian elephant housed at Dickerson Park Zoo, with
intermittent access to females (n ¼ 1–5) over a 12-year period. Behavior and signs
of musth also were recorded daily by elephant department staff. Musth indicators
(temporal gland secretion, aggression, urine dribbling) increased with musth but not
prior to it. We confirmed that temporal gland secretion was a better indicator of
behavioral musth than urine dribbling. Hormones concentrations increased as musth
approached, and presumably initiated musth indicators, but variability was high.
Therefore, these hormones cannot be used to predict the onset of musth in this
individual. Rather, the free/total testosterone ratio was a good indication of the
60-day pre-musth period. In addition, testosterone production and musth indicators
increased in intensity when a young bull at the zoo started entering musth.

Corresponding Author:
Tom Tomasi, Department of Biology, Missouri State University, Springfield, MO 65897, USA.
Email: tomtomasi@missouristate.edu
2 Psychological Reports 0(0)

Keywords
Elephant, testosterone, musth, behavior, estradiol, aggression, human/animal
interactions, well-being, welfare, distress

Introduction
The Asian elephant (Elephas maximus) is in danger of extinction in the wild, and
logistical issues limit breeding in captivity. Few mating attempts result in con-
ception, and the mortality rate of babies is high. However, captive breeding
programs may be the only way to maintain the remaining genetic diversity in
the gene pool and prevent extinction of this species. Understanding the behavior
and physiology of elephants is critical.
Proper management and handling of male elephants offer not only practical
problems, such as housing and feeding, but also the possibility of physical injury
to staff and other elephants. This is not only due to their size and strength but
also to highly aggressive behavior displayed during a temporary state that is
unique to bull elephants, called ‘‘musth,’’ a suite of behavioral and/or physio-
logical manifestations (Brannian, Griffin, & Teranova, 1989; Hall-Martin, 1987;
Poole, 1987, 1989a). Other characteristics of musth include enlargement and
draining of the temporal glands, and frequent or constant urination
(Eisenberg, McKay, & Jainudeen, 1971; Poole & Moss, 1981; Vidya &
Sukumar, 2005; Yon, Chen, Moran, & Lasley, 2008).
In the wild, elephant herds are primarily composed of related females and their
young. Post-pubertal males exist largely in a different social structure, with bach-
elor herds of young males and solitary dominant males. Once a male has reached
a sufficient maturity, he periodically experience episodes of musth. It has been
suggested that African elephant (Loxodonta africana) males enter musth concur-
rent with estrus in females (Poole, 1989b), presumably to advertize dominance
(mostly to other males) (Chelliah & Sukumar, 2013; Hollister-Smith, Alberts, &
Rasmussen, 2008; Poole, 1987). However, the onset of musth is not dependent on
the presence of estrous females. Although musth is associated with mate guarding
and successful mating in the wild, and presumably increases reproductive success,
it is not necessary for the male to be in musth to impregnate a female (Eisenberg
et al., 1971; Niemuller & Liptrap, 1991; Poole, 1989b), which also suggests that
this is related to male–male behavior. The duration and severity of musth vary,
and aggression can be unexpected and unpredictable, making the captive situ-
ation even more difficult (Hess, Schmidt, & Schmidt, 1983).
In captivity, elephants are often maintained without the benefits of the
normal herd structure and opportunities to learn ‘‘acceptable’’ behaviors
(Poole et al., 1997; Sukumar, Krishnamurthy, Wemmer, & Rodden, 1997).
Providing an environment that encourages natural behaviors can reduce this
stress (Duer, 2009; Young, 1997). In addition, elephants are often moved
from institution to institution without regard to social structure, leading to
Duer et al. 3

unrest in both herds. In a natural herd, older elephants train the young using
both negative and positive reinforcement (Chelliah & Sukumar, 2013), but ele-
phant handers have found that positive reinforcement is safer and more effective
(Olson, 2004). Because of the multiple daily interactions with people, the ele-
phant handlers are often considered herd members by females and young, often
high in the dominance hierarchy. In contrast, adult males (which would be more
solitary in nature) usually consider their handlers as competitors for the dom-
inant position in the hierarchy. When the males are in musth, and their aggres-
sive behaviors escalate, they can hurt themselves and those around them, and
great care must be taken when working with these animals.
In captive situations, historically, many elephant handlers have been attacked
and often killed by elephants (Sukumar, 1991; Taya, 1993). For this reason, in the
1980s, the American Association of Zoological Parks and Aquariums (AZA)
established three protocols for human contact with elephants (Kane,
Forthman, & Hancocks, 2005): free contact, protected contact (separated from
the elephant by a barrier), and no contact. Regardless of which method is used,
however, the animals need regular care and some risk is involved. Because of
current AZA guidelines for captive elephant care (AZA, 2012), free contact is
rarely used, so handlers are not considered part of the herd, changing the rela-
tionship between elephant and handler. While animal health procedures can still
be performed under protected contact, this only occurs if the elephant is a willing
participant. Participation by males in musth is problematic and may only be
possible if the facility has an elephant-restraint device.
Hormonal influences on behaviors are a common focus of behavioral research
(Jabbour, Asher, Smith, & Morrow, 1992; Schulte & Rasmussen, 1999; Vidya &
Sukumar, 2005). Musth in Asian and African elephants is associated with a
marked increase in total testosterone (TT) (or metabolite thereof) measured in
urine (Rasmussen, Buss, Hess, & Schmidt, 1984), serum (Brown et al., 2007;
Cooper et al., 1990; Hess et al., 1983), feces (Ganswindt, Heistermann, &
Hodges, 2005; Ganswindt et al., 2010), and temporal gland secretions (TGSs)
(Rasmussen et al., 1984, 1990). Free (not bound to carrier protein) testosterone
(FT) is believed to mirror total concentrations, but no data are available in ele-
phants. Only free hormone molecules can diffuse out of the blood and bind to
receptors on target cells which carry out the hormone’s function. Because the
amount of carrier proteins can vary with time, the amount of bound testosterone
will also vary without changing the amount of FT (i.e., biological activity). Based
on these two arguments, measures of FT could be a better indicator of testoster-
one activity than TT. The FT/TT ratio varies reflecting the body’s production
and use of testosterone, the number of carrier proteins, and availability of recep-
tor sites at target cells (Carden et al., 1998; McEwen, 1980). In both Asian and
African elephants, serum concentrations of estradiol are significantly elevated
during musth and exhibit a significant positive correlation with serum testoster-
one and aggressive behavior (Turczynski, 1993).
4 Psychological Reports 0(0)

The number of adult male Asian elephants in North American zoos is extre-
mely low. This manuscript presents a longitudinal case study containing both
endocrine and behavioral data on one captive male that spans almost 12 years
and includes over 350 blood samples. Case studies allow for a detailed analysis on
an individual, in ways that cannot be duplicated by other study designs. Because it
is not feasible to conduct such a study with wild elephants, the zoo setting was the
only alternative. However, the AZA recommends great care be used when taking
blood from males in musth. These data represent a comprehensive and unique
view of musth in Asian elephants, which may never be duplicated, due to man-
agement and time constraints to collect blood samples from males in musth.
We investigated the relationship between reproductive behaviors and hor-
mone concentrations, and how these were linked to the onset of musth. No
previous papers combine both of these data sets collected simultaneously over
a long period of time. We hypothesized that an increase in serum FT, TT, and
estradiol would parallel the onset of musth and its associated indicators (behav-
iors and overt physiological signs). Identification of a reliable musth predictor
would allow for the separation of the male from other elephants, and putting
into effect different handling procedures for the safety of the handlers. Another
objective was to determine whether the reproductive status of female elephants
(Brown, 2000; Carden et al., 1998) influenced the musth behavior of this bull.
Based upon what has been documented in African elephants (Rasmussen,
Ganswindt, Douglas-Hamilton, & Vollrath, 2008), we hypothesized that females
in estrus would induce and/or intensify measured musth behaviors. The findings
of this study may inform management and staff to enable them to more accur-
ately predict impending musth in bull elephants, thus ensuring the safety of
human handlers.

Method
Animal information and blood sample collection
The study subject was an Asian male (Onyx) who was wild-caught at a very
young age and was a performing elephant before being moved to Dickerson
Park Zoo (DPZ; Springfield, MO; 1978) when he first entered musth (age 16).
For this study, his age ranged from 23 to 35 years of age. He was housed with a
juvenile male and 2–8 adult females plus offspring during the study period
(August 1985 to February 1997). The elephants were fed herbivore pellets, vita-
min-mineral supplement, and brome hay ad libitum and given water twice a day.
They were allowed indoor/outdoor access when temperatures were >0 C, and
kept inside when it was colder.
Blood was collected daily or weekly from 1985 to 1989 for another study
(Turczynski, 1993), and weekly from 1993 to 1997 (Duer, 2004) except when this
could not be done safely due to the violent nature of the bull elephant (n ¼ 357).
Duer et al. 5

Since 1990, blood collection was performed while the elephant was in a mech-
anical restraint device (‘‘crush’’). In total, the bull experienced 27 musth epi-
sodes, 16 of which had accompanying blood samples (n ¼ 136). Blood was
collected (4  5 mL tubes) from an ear vein, held at room temperature to clot,
and then centrifuged for 15 minutes at low r/min. The separated serum was
stored at 20 C. Permission to conduct this study was granted by DPZ (a sub-
sidiary of the City of Springfield, MO), and the protocol was approved by the
DPZ research committee.

Hormone analyses
Concentrations of serum TT were measured in triplicate by a radioimmunoassay
(RIA) (Diagnostic Products Corporation (DPC), Los Angeles, CA: Testosterone
Coat-A-Count (CAC), TKTT2) validated for elephants (Duer, Carden, Schmitt,
& Tomasi, 2002). Standards ranged from 0.2 to 16.0 ng/mL, and the assay fol-
lowed the manufacturer protocol with the exception of a fivefold dilution of
male serum using the 0.0 calibrator found within the kit.
FT serum concentrations were similarly measured in triplicate (DPC: FT
CAC, TKTF2). Kit directions were adhered to with the exception of a fivefold
serum dilution as for TT. The RIA (standards ranged from 1.0 to 50.0 pg/mL)
was validated for Asian elephants by demonstrating assay linearity when non-
musth serum was diluted with 0.0 calibrator to 5%, 10%, and 50%, and musth
serum was diluted to 1%, 5%, 10%, and 50%. These samples were assayed with
results of 0.903, 2.195, and 11.256 pg/mL for non-musth serum, and 1.446,
7.758, 17.007, and 95.963 pg/mL for musth serum, respectively.
Estradiol (E) was also measured utilizing RIA (DPC Estradiol CAC,
TKE25). The kit’s protocol was used with the following exceptions: (1) half
(0.5 ml) of the recommended volume of 125I estradiol solution was used; (2)
standards of 150, 500, 1800, and 3600 pg/mL were not used; (3) the 0.0 calibrator
was used to make three lower standards: 10, 5, and 2 pg/mL; and (4) 200 ml of
serum was used instead of 100 ml. Serum pool volumes of 50, 100, and 200 ml
were analyzed with resulting values of non-musth values of 0.004, 0.007, and
0.011 pg/mL respectively, and musth values of 0.006, 0.010, and 0.018 pg/mL,
respectively.

Behavioral observations
DPZ elephant keepers documented physical condition and observed behaviors
during their daily care of the elephant (n ¼ 2343 days). The behavioral data were
collected from 1993 to 1997 using a daily data sheet (‘‘ethogram’’) completed by
the elephant keepers (see supplemental document). It was constructed based on
previously reported behaviors and physical conditions related to male elephants
in musth, and observations of behavior by the elephant staff at DPZ (Duer,
6 Psychological Reports 0(0)

Table 1. Subjective scoring rubrics for musth condition/index used by elephant keepers on
daily ethogram (see supplemental photo of moderate temporal gland secretion).

Scores

Index 1 2 3 4

Temporal gland None Slight (3–12 cm Moderate (12– Heavy (dripping

secretion wet spot) 25 cm wet off face)
spot)
Aggression None Irritable (obvi- Aggressive Very aggressive
ous agitation) (agonistic (agonistic
displays) behaviors)
Crush entry – Yes Refusal –
Urine dribbling None Moderate (wet Heavy (visible –
inner legs) drainage)
Breeding behavior No interest Mating interest Mounting Confirmed
penetration

2004). This data sheet was divided into three major sections: behavior during the
daily routine, aggressive behavior, and physical condition. Behaviors recorded
were aggression level, breeding, and entry into (or refusal to enter) the crush.
Other musth ‘‘indicators’’ recorded were TGS and urine dribbling (UD). At the
onset of the study, the behaviors and indicators associated with musth were
operationally defined by the DPZ elephant committee (researchers, keepers,
administrators, veterinarian). Data for each condition or behavior (‘‘indices’’)
were classified into 2–4 categories, for statistical comparisons of endocrine data
(Table 1). To maintain consistency in reporting, the elephant keepers developed
detailed scoring rubrics. We also recorded when females were in estrus, based on
serum progesterone concentrations (Carden et al., 1998).

Data analyses
For this case study (this bull), the empirically derived operational definition of
musth is five consecutive days of TGS: the first of these five days was considered
the first day of musth. Similarly, we considered the bull to be in non-musth after
five consecutive days with no TGS; the first of these days was considered the first
day of non-musth. Because elephant managers are concerned about anticipating
the onset of musth, we considered this first day of non-musth as the start of the
then next musth cycle (non-musth, pre-musth, musth), rather than calling it
post-musth.
Endocrine and behavioral data were divided into five time periods for statis-
tical comparison: (1) musth, (2) the seven days prior to musth onset, (3) 8–14
days prior to musth, (4) 15–60 days prior to musth (the 60 days prior to musth
Duer et al. 7

are collectively called ‘‘pre-musth’’), and (5) more than 60 days prior to musth
(called ‘‘non-musth’’ herein). One-way analysis of variance (ANOVA) was used
to test for differences among the time periods. When the ANOVA was signifi-
cant, a Tukey’s test was used to assess pairwise comparisons. Using the same five
time periods, a chi-square test was used to determine whether a temporal rela-
tionship existed between a female entering estrus and the bull entering musth.
Endocrine data were also sorted by (i) the intensity level of the musth indi-
cators, (ii) whether breeding behavior (BB) was observed, and (iii) the number of
females in estrus. In addition, to assess the effect of the juvenile male maturing
and entering musth on his own (1993–1997), data for intensity of musth indica-
tors in this four-year period were separated from the previous four-year intervals
(1 March to 28 February). Data for daily musth intensity values were also sorted
based on the number of females in estrus. Each data set was compared with a
one-way ANOVA and Tukey’s test. For all tests, a ¼ .05 was used for significant
differences. All data are presented as means SE.

Results
Blood samples for hormone analysis (1985–1997) and observations by elephant
handlers (1981–1997) combine to create the most extensively documented profile
of an individual captive male Asian elephant yet available, and together support
the episodic nature of musth (Figure 1). Over this period, blood collection fre-
quency was dependent on ongoing research, and observation of musth indica-
tors after August 1995 was more formalized.

Musth timeline
Significant increases in serum TT (ng/mL) and FT (pg/mL) were found with the
onset of musth (p < .001; Table 2), although there was a considerable variation
in hormone concentrations within each phase of the musth cycle. This pattern of
increasing means is also seen within each individual musth cycle (Figure 2).
There was also an increase in the FT/TT ratio (p < .001) with significantly
higher ratio during musth and the preceding 60 days than during non-musth
(Table 2).
Estradiol (E; pg/mL) concentrations also increased as musth approached such
that levels in musth were significantly higher than in non-musth and >15 days
prior to musth (p < .001; Table 2). The ratio of E to TT declined steadily as
musth approaches (p < .001; Table 2), and as with E concentrations, the E/TT
ratios during musth were different from non-musth and >15 days prior to
musth.
Intensities of TGS, aggression, and UD increased during musth compared to
any of the pre-musth periods (p < .001; Table 2). Intense aggression was readily
elicited during musth and could be focused on any object/person that attracted
8 Psychological Reports 0(0)

Figure 1. Intensity of musth indicators over a 16-year period for one male Asian elephant,
and his testosterone levels (ng/mL). Data over this time were taken primarily associated
with two research projects (Duer, 2004; Turczynski, 1993).

the bull’s attention. This kind of aggressive behavior was rare during non-musth.
TGS and aggression gradually increased prior to musth, but UD was only
observed during the musth episode.
The average number of females in estrus varied among the phases of the
musth cycle (p < .001; Table 2). This value increased gradually from day 60
to musth, but more females were in estrus during non-musth than any other time
period. Proportionately, females entered estrus as often when the bull was in
non-musth, in pre-musth (60 days prior to musth onset), or in musth (2;
p ¼ .61).
The BB index showed a matching pattern, with a higher BB during non-
musth than when in musth (p < .001; Table 2). Refusals to enter and aggression
towards the crush were more common when in musth than any other time period
(Table 2), and more frequently in the morning than the afternoon (Figure 3).

Hormone/behavior correlates
Concentrations of TT and FT increased with the level of TGS, the level of
aggression, and the extent of UD (Figure 3). The same pattern was observed
Table 2. Endocrine measures and musth indicators throughout the musth cycle for one male elephant.
Duer et al.

Measure Units Non-musth 60 days 14 days 7 days Musth p

a a,b b,c b,c c
Total testosterone ng/mL 10.2  1.7 20.1  +2.3 31.8  6.4 36.0  6.3 48.8  2.7 <0.001
(73) (70) (16) (22) (176)
Free testosterone pg/mL 34.8  6.7a 97.1  11.4a,b 129.0  24.5a,b,c 173.5  27.3b,c 199.5  13.1c <0.001
(73) (70) (16) (22) (176)
Free/total testosterone NA 0.0026  0.0002a 0.0042  0.0003b 0.0039  0.0004b 0.0047  0.0004b 0.0037  0.0001b <0.001
ratio (73) (70) (16) (22) (176)
Estradiol pg/mL 4.5  0.0001a 4.6  0.0003a 5.0  0.0003a,b 5.7  0.0005a,b 5.6  0.0001b <0.001
(57) (22) (6) (7) (101)
Estradiol/total testoster- NA 0.0022  0.0003a 0.0016  0.0005a 0.0011  0.0006a,b 0.0007  0.0004a,b 0.0003  0.0000b <0.001
one ratio (57) (22) (6) (7) (101)
Temporal gland secretion 1–4 scale 1.05  0.01a 1.04  0.01a 1.08  0.03a 1.15  0.04a 2.76  0.03b <0.001
(592) (410) (114) (131) (1091)
Aggression 1–4 scale 1.19  0.02a 1.19  0.02a 1.23  0.05a 1.40  0.06a 2.19  0.03b <0.001
(591) (410) (114) (131) (1091)
Urine dribbling 1–3 scale 1.00  0.00a 1.00  0.00a 1.00  0.00a 1.02  0.01a 1.50  0.02b < 0.001
(592) (410) (114) (131) (1091)
Females in estrus No. of females 0.49  0.03a 0.18  0.02b 0.21  0.04b 0.31  0.05b,c 0.40  0.02c <0.001
(592) (410) (114) (131) (1091)
Breeding behavior 1–4 scale 1.75  0.04a 1.44  0.04b 1.43  0.06b,c 1.60  0.07a,b,c 1.59  0.03c <0.001
(591) (410) (114) (131) (1091)
AM Crush refusal % 7.6 13.6 13 17.8 29.5
PM Crush refusal % 5.6 5.2 6.2 10.0 23.5

Data (means  SE, and n) are broken down into the musth period itself (defined by temporal gland secretion), the week preceding the onset of musth (days 7 to
1), the time periods of 8–14 days and 15–60 days prior to musth onset, and ‘‘non-musth’’ (>60 days prior to musth). On the same row, superscripts sharing
letters are not different (Tukey’s test; p < .05).
9
10 Psychological Reports 0(0)

Figure 2. Serum testosterone levels during individual musth cycles. The first day of musth
is designated as day 0, and all days beyond that are in musth. Data collected after musth
ended were consider pre-musth for the next musth period and are shown as negative days.

Figure 3. Serum concentrations (mean  SE) of testosterone (total and free) and estradiol
(total) in a captive male Asian elephant at various intensities of three musth indicator: tem-
poral gland secretion, aggression, and urine dribbling. Also shown are testosterone concen-
trations (total and free) on days when he was given access to a hydraulic restraint device
(crush), sorted by whether he entered or refused to enter it. Data for each hormone were
analyzed with a one-way ANOVA, and shared letters above bars indicate no difference
among means (Tukey’s test).
Duer et al. 11

for E, but it was not as pronounced (Figure 3). TT and FT levels were elevated
when multiple females were in estrus (relative to zero females), but not when
only one female was in estrus (Figure 4). Similarly, musth indicators were more
intense when multiple females were in estrus but not with just one estrus female
(Figure 5). E levels in the male did not change with females in estrus. FT and E
levels were lower on days when BB was observed (Figure 4); although this trend
was evident in TT, the difference was not significant (p ¼ .09). TT and FT levels

Figure 4. Serum concentrations (mean  SE) of testosterone (total and free) and estradiol
(total) in a captive male Asian elephant given access to females, sorted by whether or not
breeding behavior was observed, and also sorted by the number of estrus females present.
Data for each hormone were analyzed with a t test for breeding behavior, or with a one-
way ANOVA, and shared letters above bars indicate no difference among means (Tukey’s
test).
12 Psychological Reports 0(0)

Figure 5. Intensity of musth indicators during days with different numbers of females in
estrus. Data for each musth indicator were analyzed with a one-way ANOVA, and shared
letters above/below symbols indicate no difference among means (Tukey’s test).

were higher on days when afternoon crush entry was refused (Figure 3). This
was not the case when opportunities to enter the crush were given in the morn-
ing, but the endocrine pattern was similar. Data on E levels were not sufficient to
compare for this behavior.
A juvenile male housed at the zoo did not enter musth until the last three
years of the study, and prior to this, we do not think that his presence altered the
musth cycling of our focus male. Once the juvenile entered his first musth (age
16), our focal male entered musth more often and more sporadically. During the
final four years of the study, when we have behavioral data for about 95% of the
days, he was in musth for 51% of these days. In the preceding three four-year
periods, he was in musth for 27%–46% of the days that records were kept. These
are likely to be over-estimates, as behaviors were recorded less than half of the
days in the time period (6%–44% of the possible days), and days with no record
are more likely to be days when no aggression or sign of musth was observed. In
addition, his intensity of aggression, TGS, and UD while in musth was greater in
the last four-year period (p < .001), when the younger bull was also entering
musth (Figure 6) than in any four-year period prior to the younger bull’s
musth. Consistent with this, our focal male had higher levels of TT and FT
concentrations while in musth during this last four-year period than an earlier
Duer et al. 13

Figure 6. Serum concentrations (mean  SE) of testosterone (total and free) (top), and
intensity of musth indicators (bottom) in our focus male, before and after the first musth
episode (late in 1993) of the juvenile male housed at Dickerson Park Zoo. Since we have
four years of observation when both males were entering musth, the period prior to this
was also divided into four-year increments. Data for each hormone and musth indicator
were analyzed with a one-way ANOVA, and shared letters above/below bars indicate no
difference among means (Tukey’s test).

period when we had sufficient blood samples for a meaningful comparison

(p < .001; Figure 6).

Discussion
The a priori operational definitions for musth were five continuous days of TGS,
and for non-musth was five continuous days of no TGS. Although these defin-
itions may not work for all male elephants, musth defined in this way correlates
well with hormone concentrations, aggression, and UD (other indicators of
musth; Table 2; Figure 1) in our focal elephant. Although the TGS was not
continuous, most breaks were less than five days, so that ‘‘musth’’ was more
14 Psychological Reports 0(0)

Figure 7. Data from Figure 1 expanded to show five months in 1993. Shaded area in the
TGS plot indicates that this bull is in musth based on the operational definition used for
this study. Note that temporal gland secretion intensity more closely mirror aggression
intensity than does urine dribbling for this bull.

continuous than TGS and we avoided having hundreds of very short musth
periods.
Previous work on musth in African elephants (Ganswindt et al., 2005;
Ganswindt, Rasmussen, Heistermann, & Hodges, 2005; Hollister-Smith et al.,
2008; Poole, 1987) suggests that UD may be a better indicator of musth than
TGS. However, this was not the case with our Asian bull. We did not find that
UD occurred each time aggression increased, and intensity of TGS matched
aggression episodes much better (Figure 7). In general, UD occurred only
when the periods of aggression and TGS were longer or more intense, which
also coincided with when the younger bull started entering musth (Figure 6).
This may be attributed to an individual or species difference. Ganswindt,
Heistermann, et al. (2005), Ganswindt, Rasmussen, et al. (2005), Ganswindt
et al. (2010), and Cooper et al. (1990) highlight endocrine and musth difference
between the two elephant species. Since one purpose of this study is to look for
pre-musth markers to predict aggression in captive elephants (for the safety of
the elephant handlers), TGS should remain a primary musth indicator in Asian
elephants.
Duer et al. 15

Aggression was also more variable than TGS because other events can trigger
aggressive behavior, such as novel equipment (construction) or people he did not
know, strange elephants at the zoo, operation of the hydraulic doors, etc.
Refusal to enter the ‘‘crush’’ or follow daily routines was included in this
aggression.
Even though TGS was the most obvious physical sign of musth in our focal
bull, it has been suggested that TGS and musth may not be strictly associated in
African elephants (Eisenberg et al., 1971; Poole & Moss, 1981). In both Asian
and African bulls, TGS can occur at times other than during musth. In our
study, these seldom reached the five-day requirement to qualify as musth.
Female African elephants also experience TGS (Kahl & Amstrong, 2002), and
to a lesser degree, Asian females may as well, but this would not be sufficient to
meet our definition of musth. Therefore, our operational definition was useful in
separating musth and non-musth for this Asian elephant and supports the
conclusions of Yon et al. (2008), but would not be appropriate for African
elephants. Occasionally, our elephant fluctuated between musth and non-
musth. Some of these short interruptions of musth appeared to be triggered
when physically subdued in the crush for collection of semen or routine
maintenance.

Endocrinology throughout the musth cycle

TT increased with the approach of musth, starting at least two weeks before
TGS began (Table 2). However, there was a lot of variation in this measure,
which is consistent with previous studies (Brown et al., 2007; Cooper et al., 1990;
Ganswindt et al., 2010; Hess et al., 1983; Niemuller & Liptrap, 1991). Clearly, a
single blood sample and TT measurement would not be sufficient to predict
upcoming musth. The time of day each blood sample was taken can also con-
tribute to the variation in endocrine values because of pulsatile secretion from
glands and circadian rhythms. To reduce this component of variation, samples
were taken during the morning hours whenever possible. The TT mean we
report here during musth is probably an underestimate. Keepers were not able
to collect blood on the days when the bull was most aggressive, so it is likely we
missed some of the highest concentrations.
FT concentrations tracked the rise and fall of TT and there was little change
in the FT/TT ratio 60 days prior to and in musth. These suggest that binding
proteins in serum increased proportionally as musth approached, and the
increasing FT indicates a gradually increasing set point for testosterone in the
hypothalamus. Since this FT/TT ratio was lower when more than 60 days prior
to musth (Table 2), regular measurements might provide a better long-term
predictor of musth than TT or FT alone.
Other androgens may be even better predictors, especially when expressed
relative to testosterone levels. Higher testosterone/dihydrotestosterone ratios
16 Psychological Reports 0(0)

were found during musth, and significantly higher ratios were found during the
heaviest part of musth (Dickerman, Zachariah, Fouraker, & McConathy, 1997;
Rasmussen et al., 1984; Rasmussen, Hess, & Haight, 1990). Similarly, a higher
testosterone/androstenedione ratio was found during musth than during non-
musth (Brannian et al., 1989; Niemuller & Liptrap, 1991). Increases in testos-
terone concentrations follow pulsatile increases in plasma luteinizing hormone
(LH) concentrations (Brannian et al., 1989; Niemuller & Liptrap, 1991), con-
sistent with the general observation that LH stimulates testicular steroidogen-
esis. LH levels begin to rise four weeks prior to musth and then decrease
immediately after the onset of TGS (Turczynski, 1993). In African elephants,
a positive correlation between musth-related aggression and androgenic metab-
olite concentrations (epiandrosterone) has been demonstrated in fecal samples
(Ganswindt, Heistermann, et al., 2005; Ganswindt, Rasmussen, et al., 2005;
Rasmussen et al., 2008).
Turczynski (1993) reported that E concentrations in Asian males during non-
musth (8.7 3.2 pg/mL) were similar to concentrations in female Asian ele-
phants (Brown, Walker, & Moeller, 2004; Taya et al., 1991), but these were
then elevated during musth (19.5  10.7 pg/mL). Results of our study provide
values at the low end of this range (4–7 pg/mL) but with a similar pattern of TT
changes through the musth cycle (Table 2). The decline in the ratio of E/TT as
musth approached demonstrates that TT increases more than E increases with
onset of musth (Table 2). The source of E in male Asian elephants is not known
but a major source is presumed to be the testicles via aromatization of androgens
since E is synthesized from T (or other androgens) by enzymes generally found
in the testes (Gorbman, Dickhoff, Vigna, Clark, & Ralph, 1983). Therefore, it is
not surprising that E increased in parallel with T values. It is not known, how-
ever, whether this E increase is solely a consequence of increasing T production,
or whether it performs a physiological role in musth.

Indicators of the musth cycle

For our focal male, TGS and aggression (including refusal to enter the crush)
displayed the same pattern of increasing intensity prior to the onset of musth,
peaking during musth, and decreasing after the musth episode. This pattern was
consistent with previous studies on both elephant species (Cooper et al., 1990;
Dickerman et al., 1997; Lincoln & Ratnasooriya, 1996; Rasmussen & Schulte,
1998). The higher number of crush refusals in the morning may be because
refusal in the evening resulted in going without food pellets for a longer
period (all night vs. 8 hours).
Once the younger resident bull started entering musth, our focal bull (16 years
older) experienced more days in musth, had higher levels of testosterone when in
musth, and demonstrated more intense musth indicators (Figures 1 and 6). To
our knowledge, this response has never been quantified, or even described
Duer et al. 17

qualitatively. Although the mature bull had lived with the calf for 14 years prior
to the younger bull entering musth, these changes suggest that he apparently
now considered the young bull to be a rival male. This supports the conclusions
of Poole (1987), Cooper et al. (1990), and Rasmussen and Schulte (1998) that
musth is primarily a male–male interaction, used to advertize physiological state
and social order. Understanding this interaction is important to inform man-
agement decisions where two bulls are housed at the same institution.
There was a complex relationship between musth and number of females in
estrus, with testosterone and musth indicators elevated when more than one
female was in estrus (Figures 4 and 5). This observation is complicated because
having more than one female in estrus was more likely to occur when two bulls
were in musth, as more females were brought in for breeding purposes during
this time. However, for three reasons, we believe that the increase in musth
indicators in our focal male at this time was to communicate dominance over
the younger bull entering musth, rather than due to the additional females in
estrus. First, there was no increase in musth indicators with just one female in
estrus, which occurred more often prior to the younger male entering his first
musth period. Second, a male Asian elephant can impregnate a female without
being in musth (Eisenberg et al., 1971; Hall-Martin, 1987; Niemuller & Liptrap,
1991; Poole, 1987, 1989b). Third, if females in estrus were a physiological stimu-
lus for musth, then estrus would precede musth onset. However, based upon the
proportion of females entering estrus throughout the musth cycle, there was no
relationship between the onset of musth and the onset of estrus.
Because our elephant was not allowed free access to the females, interpreting
these BB data must be done with caution. His contact with females was deter-
mined by management decisions related to the breeding program (AZA species
survival plan), and without regard to musth. Therefore, opportunities to breed
were random in relation to our stages of the musth cycle (access to females was
allowed in 35%–50% of the days in each stage). We conclude that our data
support previous observations that males will breed with females whether in
musth or not.

Conclusion
We affirmed that TGS is the best indicator of musth for this captive Asian bull,
as it associates with aggression better than does UD. This is in contrast to
studies on wild African elephants where UD is proposed to be a better indicator.
We hypothesized that musth and its associated indicators would positively
correlate with testosterone levels; this was clearly demonstrated. However, tes-
tosterone measurement (by itself) is not a good predictor of musth for manage-
ment purposes because there is a lot of variation in the hormone concentrations,
and the increase was gradual prior to musth. It would be interesting to examine
how dihydrotestosterone and androstenedione changed during the musth cycle,
18 Psychological Reports 0(0)

as these may change more consistently and dramatically, and be a better pre-
dictor for elephant managers. We anticipated that E might have a strong influ-
ence on musth behavior, but it was relatively level over the musth cycle. Either it
is not having much effect on musth or the influence is exerted via a change in
cellular response to the E.
We recommend that elephant-management decisions, such as isolating and
confining the male in musth, and strictly adhering to the AZA no-contact pro-
cedures, continue to be made based mostly on behavior of the bull and asso-
ciated musth indicators. If institutions do not have the facilities to isolate/
confine them and the willingness to follow safety recommendations, it would
not be prudent to house musth-age bulls. In Asian countries where elephants
work alongside humans, bulls in musth are given large doses of cocaine to
reduce aggressive behaviors (Sukumar et al., 1997). However, maintaining
them on medication or castration (chemically) is not considered humane in
the United States.
Depending upon endocrine markers to identify impending musth is not rec-
ommended, based on our data. The gradual increase in testosterone levels is
masked by the variability, which may be due to the episodic release of steroid
hormones, and where in the release cycle the bull is when the blood is collected.
Among other factors that might influence onset, duration, and termination of
musth in captive elephants, a change in the social structure or reproductive
status of an elephant (such as a younger bull entering musth for the first
time), should alert management to increase safety precautions.

Acknowledgments
The authors thank the Dickerson Park Zoo management, staff, and the elephant com-
mittee, especially Melissa Carden and John Bradford for their support. We also appre-
ciate the funding assistance from Missouri State University and the Department of
Biology, encouragement from Dr. ‘‘Bets’’ Rasmussen, and manuscript suggestions from
Janine Brown. The authors dedicate this paper to John Bradford who dedicated his life to
the conservation and understanding of Asian elephants.

Declaration of Conflicting Interests

The authors declare no potential conflicts of interest with respect to the research, author-
ship, and/or publications of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, author-
ship, and/or publication of this article: This work received internal funding from Missouri
State University.

Author Contributions
CD and TT contributed equally to this work.
Duer et al. 19

References
AZA. (2012). AZA standards for elephant management and care. Retrieved from https://
www.aza.org/uploadedFiles/Accreditation/AZA%20Standards%20for%20Elephant
%20Management%20and%20Care.pdf
Brannian, J. D., Griffin, F., & Teranova, P. F. (1989). Urinary androstenedione and
luteinizing hormone concentrations during musth in a mature African elephant. Zoo
Biology, 8, 165–170.
Brown, J. L. (2000). Reproductive endocrine monitoring of elephants: An essential tool
for assisting captive management. Zoo Biology, 19, 347–367.
Brown, J. L., Somerville, M., Riddle, H. S., Keele, M., Duer, C. K., & Freeman, E. W.
(2007). Comparative endocrinology of testicular, adrenal, and thyroid function in
captive Asian and African elephant bulls. General and Comparative Endocrinology,
151, 153–162.
Brown, J. L., Walker, S. L., & Moeller, T. (2004). Comparative endocrinology of cycling
and non-cycling Asian (Elephas maximus) and African (Loxodonta africana) elephants.
General and Comparative Endocrinology, 136, 360–370.
Carden, M., Schmitt, D., Tomasi, T., Bradford, J., Moll, D., & Brown, J. (1998). Utility
of serum progesterone and prolactin analysis for assessing reproductive status in the
Asian elephant (Elephas maximus). Animal Reproduction Science, 53, 133–142.
Chelliah, K., & Sukumar, R. (2013). The role of tusks, musth and body size in male-male
competition among Asian elephants, Elephas maximus. Animal Behaviour, 86,
1207–1214.
Cooper, K. A., Harder, J. D., Clawson, D. H., Frederick, D. L., Lodge, G. A., Peachey,
H. C., . . . Winstel, D. P. (1990). Serum testosterone and musth in captive male African
and Asian elephants. Zoo Biology, 9, 297–306.
Dickerman, R. D., Zachariah, N. Y., Fouraker, M., & McConathy, W. J. (1997).
Neuroendocrine-associated behavioral patterns in the male Asian elephant (Elephas
maximus). Physiology & Behavior, 61, 771–773.
Duer, C. K. (2004). Reproductive endocrinology and musth-behavior of a captive Asian
elephant (Elephas maximus) (Master’s thesis). Missouri State University,
Springfield, MO.
Duer, C. K. (2009). Function of play in behavioral enrichment of captive Asian elephants
(Elephas maximum) (Doctoral dissertation, Oklahoma State University, Stillwater,
OK).
Duer, C. K., Carden, M., Schmitt, D., & Tomasi, T. (2002). Utility of maternal serum
total testosterone analysis for fetal gender determination in Asian elephants (Elephas
maximus). Animal Reproduction Science, 69, 47–52.
Eisenberg, J. F., McKay, G. M., & Jainudeen, M. R. (1971). Reproductive behavior of
the Asiatic elephant (Elephas maximus). Behaviour, 38, 193–224.
Ganswindt, A., Heistermann, M., & Hodges, K. (2005). Physical, physiological, and
behavioral correlates of musth in captive African elephants (Loxodonta africana).
Physiological and Biochemical Zoology, 78, 505–514.
Ganswindt, A., Muenscher, S., Henley, M., Henley, S., Heistermann, M., Palme, R., . . .
Bertschinger, H. (2010). Endocrine correlates of musth and the impact of ecological
and social factors in free-ranging African elephants (Loxodonta africana). Hormones
and Behavior, 57, 506–514.
20 Psychological Reports 0(0)

Ganswindt, A., Rasmussen, H. B., Heistermann, M., & Hodges, J. K. (2005). The sexu-
ally active states of free-ranging male African elephants (Loxodonta africana):
Defining musth and non-musth using endocrinology, physical signals, and behavior.
Hormones and Behavior, 47, 83–91.
Gorbman, A., Dickhoff, W. W., Vigna, S. R., Clark, N. B., & Ralph, C. L. (1983).
Comparative endocrinology. New York, NY: John Wiley & Sons.
Hall-Martin, A. J. (1987). Role of musth in the reproductive strategy of the African
elephant (Loxodonta africana). South African Journal of Science, 83, 616–621.
Hess, D. L., Schmidt, A. M., & Schmidt, M. J. (1983). Reproductive cycle of the Asian
elephant (Elephas maximus) in captivity. Biology of Reproduction, 28, 767–773.
Hollister-Smith, J. A., Alberts, S. C., & Rasmussen, L. E. L. (2008). Do male African
elephants, Loxodonta africana, signal musth via urine dribbling?. Animal Behaviour,
76, 1829–1841.
Jabbour, H. N., Asher, G. N., Smith, J. F., & Morrow, C. J. (1992). Effect of proges-
terone and oestradiol benzoate on oestrous behavior and secretion of luteinizing hor-
mone in ovariectomized fallow deer Dama dama. Journal of Reproduction and Fertility,
94, 353–361.
Kahl, M., & Amstrong, B. D. (2002). Visual displays of wild African elephants during
musth. Mammalia, 66, 159–171.
Kane, L., Forthman, D., & Hancocks, D. (2005). Optimal conditions for captive ele-
phants: A report by the coalition for captive elephant well-being. Retrieved from
http://www.elephantcare.org/protodoc_files/2008/CCEWBOptimalConditionspdf.2.pdf
Lincoln, G. A., & Ratnasooriya, W. D. (1996). Testosterone secretion, musth behaviour
and social dominance in captive male Asian elephants living near the equator. Journal
of Reproduction and Fertility, 108, 107–113.
McEwen, B. S. (1980). The brain as a target organ of endocrine hormones. In D.
T. Krieger & C. Hughes (Eds.), Neuroendocrinology (pp. 33–44). Sunderland, MA:
Sinauer Associate.
Niemuller, C. A., & Liptrap, R. M. (1991). Altered androstenedione to testosterone ratios
and LH concentrations during musth in the captive male Asian elephant (Elephas
maximus). Journal of Reproduction and Fertility, 91, 131–146.
Olson, D. (2004). Elephant husbandry resource guide. Lawrence, KS: Allen Press.
Poole, J. H. (1987). Rutting behavior in African elephants: The phenomenon of musth.
Behaviour, 102, 283–316.
Poole, J. H. (1989a). Announcing intent: The aggressive state of musth in African ele-
phants. Animal Behaviour, 37, 140–152.
Poole, J. H. (1989b). Mate guarding, reproductive success and female choice in African
elephants. Animal Behaviour, 37, 842–849.
Poole, J. H., & Moss, C. J. (1981). Musth in the African elephant, Loxodonta africana.
Nature, 292, 830–831.
Poole, T. B., Taylor, V. J., Fernando, S. B. U., Ratnosooriya, W., Ratnayeke,
A., Lincoln, G. L., . . . Manatunga, A. M. V. R. (1997). Social behavior and
breeding physiology of a group of Asian elephants Elephas maximus at the
Pinnawala Elephant Orphanage, Sri Lanka. International Zoo Yearbook, 35,
297–310.
Duer et al. 21

Rasmussen, L. E., Buss, I. O., Hess, D. L., & Schmidt, M. J. (1984). Testosterone and
dihydrotestosterone concentrations in elephant serum and temporal gland secretions.
Biology of Reproduction, 30, 352–362.
Rasmussen, H. B., Ganswindt, A., Douglas-Hamilton, I., & Vollrath, F. (2008).
Endocrine and behavioral changes in male African elephants: Linking hor-
mone changes to sexual state and reproductive tactics. Hormones and Behavior, 54,
539–548.
Rasmussen, L. E. L., Hess, D. L., & Haight, J. D. (1990). Chemical analysis of temporal
gland secretions collected from an Asian bull elephant during a four-month musth
episode. Journal of Chemical Ecology, 16, 2167–2181.
Rasmussen, L. E. L., & Schulte, B. A. (1998). Chemical signals in the reproduction of
Asian (Elephas maximus) and African (Loxodonta africana) elephants. Animal
Reproduction Science, 53, 19–34.
Schulte, B. A., & Rasmussen, L. E. L. (1999). Musth, sexual selection, testosterone, and
metabolites. In R. E. Johnson, D. Muller-Schwarze & W. Sorensen (Eds.), Eighth
international symposium on chemical signals in vertebrates (pp. 383–397). Ithaca, NY:
Plenum Publishers.
Sukumar, R. (1991). The management of large mammals in relation to male strategies
and conflict with people. Biological Conservation, 55, 93–102.
Sukumar, R., Krishnamurthy, V., Wemmer, C., & Rodden, M. (1997). Demography of
captive Asian elephants (Elephas maximus) in southern India. Zoo Biology, 16,
263–272.
Taya, K. (1993). The reproductive physiology of the elephant. Reproduction and
Development, 39, 77–91.
Taya, K., Kondoh, M., Ogawa, Y. L., Nakada, K., Watanabe, G., Sasamoto, S., . . .
Narushima, E. (1991). Concentrations of progesterone, testosterone and estradiol-17B
in the serum during the estrous cycle of Asian elephants (Elephas maximus). Zoo
Biology, 10, 299–307.
Turczynski, C. J. (1993). The endocrinology of musth in the male Asian elephant (Elephas
maximus), serum estradiol, serum LH and serum, fecal and urinary testosterone
(Doctoral dissertation). Texas A&M University, Texas, USA.
Vidya, T. N. C., & Sukumar, R. (2005). Social and reproductive behaviour in elephants.
Current Science, 89, 1200–1207.
Yon, L., Chen, J., Moran, P., & Lasley, B. (2008). An analysis of the androgens of musth
in the Asian bull elephant (Elephas maximus). General and Comparative Endocrinology,
155, 109–115.
Young, R. J. (1997). Designing environmental enrichment devices around species specific
behavior. In B. Holst (Ed.), Proceedings of the second international conference on
environmental enrichment (pp. 195–204). Copenhagen, Denmark: Copenhagen Zoo.

Author Biographies
Connie K. Duer, earned her PhD at Oklahoma State University. Her research
interests are in elephants, animal communication, and behavior.
22 Psychological Reports 0(0)

Tom Tomasi, earned his PhD at University of Utah. His primary research inter-
est is in animal energetics and physiological ecology.

Charles I. Abramson, earned his PhD at Boston University. His major research
interest is in the comparative analysis of behavior.

View publication stats