Serum testosterone and musth in captive male African and Asian elephants

Testosterone concentrations in serum samples collected weekly over a 5-year period from a young adult male Asian elephant (Elephas maximus) and a young adult male African forest elephant (Loxodonta africana cyclotis) were measured by radioimmunoassay. Testosterone profiles during this maturational period were compared between the two species and related to the occurrence of musth, a recurring physiological and behavioral condition exhibited by most mature Asian, and some African, bull elephants. Musth is characterized by secretion from the bull's temporal glands, dribbling urine, and increased aggression. Serum testosterone concentrations in the Asian bull were elevated substantially between April and September each year, coincident with the presence of temporal gland secretion, urine dribbling, and aggressive behavior. Testosterone levels from April through September averaged (± SEM) 41.2 ± 2.8 ng/ml, compared to 7.6 ± 1.0 ng/ml during the rest of the year. In contrast, the testosterone profile of the African bull showed greater variation and lower levels overall, the only pattern being a tendency for levels to be lowest from November to February (avg. 6.8 ± 1.5 vs. 10.3 ± 0.8 ng/ml during the rest of the year). Temporal gland secretion and other signs of musth were first observed in this bull in 1988, at age 17. While his testosterone profile did not show a pattern comparable to that in the Asian bull, average testosterone values were significantly greater in 1988 compared to previous years. The Asian bull showed sexual attention to preovulatory (estrous) cows whether in musth or not, and exposure to estrous cows did not appear to alter the highly consistent, annual pattern of musth as evidenced in temporal gland flow.     

Endocrine Correlates of Musth in Free-Ranging Asian Elephants (Elephas maximus) Determined by Non-Invasive Faecal Steroid Hormone Metabolite Measurements

The occurrence of musth, a period of elevated levels of androgens and heightened sexual activity, has been well documented for the male Asian elephant (Elephas maximus). However, the relationship between androgen-dependent musth and adrenocortical function in this species is unclear. The current study is the first assessment of testicular and adrenocortical function in free-ranging male Asian elephants by measuring levels of testosterone (androgen) and cortisol (glucocorticoid – a physiological indicator of stress) metabolites in faeces. During musth, males expectedly showed significant elevation in faecal testosterone metabolite levels. Interestingly, glucocorticoid metabolite concentrations remained unchanged between musth and non-musth periods. This observation is contrary to that observed with wild and captive African elephant bulls and captive Asian bull elephants. Our results show that musth may not necessarily represent a stressful condition in free-ranging male Asian elephants.     

Physical, physiological, and behavioral correlates of musth in captive African elephants (Loxodonta africana)

Although musth in male African elephants (Loxodonta africana) is known to be associated with increased aggressiveness, urine dribbling (UD), temporal gland secretion (TGS), and elevated androgens, the temporal relationship between these changes has not been examined. Here, we describe the pattern of musth-related characteristics in 14 captive elephant bulls by combining long-term observations of physical and behavioral changes with physiological data on testicular and adrenal function. The length of musth periods was highly variable but according to our data set not related to age. Our data also confirm that musth is associated with elevated androgens and, in this respect, show that TGS and UD are downstream effects of this elevation, with TGS responding earlier and to lower androgen levels than UD. Because the majority of musth periods were associated with a decrease in glucocorticoid levels, our data also indicate that musth does not represent a physiological stress mediated by the hypothalamic-pituitary-adrenal axis. Furthermore, we demonstrate that the occurrence of musth is associated with increased aggression and that this is presumably androgen mediated because aggressive males had higher androgen levels. Collectively, the information generated contributes to a better understanding of what characterizes and initiates musth in captive African elephants and provides a basis for further studies designed to examine in more detail the factors regulating the intensity and duration of musth.     

The sexually active states of free-ranging male African elephants (Loxodonta africana): defining musth and non-musth using endocrinology, physical signals, and behavior

Musth in male African elephants, Loxodonta africana, is associated with increased aggressive behavior, continuous discharge of urine, copious secretions from the swollen temporal glands, and elevated androgen levels. During musth, bulls actively seek out and are preferred by estrous females although sexual activity is not restricted to the musth condition. The present study combines recently established methods of fecal hormone analysis with long-term observations on male-female associations as well as the presence and intensity of physical signals to provide a more detailed picture about the physical, physiological, and behavioral characteristics of different states of sexual activity in free-ranging African elephants. Based on quantitative shifts in individual bull association patterns, the presence of different physical signals, and significant differences in androgen levels, a total of three potential sub-categories for sexually active bulls could be established. The results demonstrate that elevations in androgen levels are only observed in sexually active animals showing temporal gland secretion and/or urine dribbling, but are not related to the age of the individual. Further, none of the sexually active states showed elevated glucocorticoid output indicating that musth does not represent an HPA-mediated stress condition. On the basis of these results, we suggest that the term "musth" should be exclusively used for the competitive state in sexually active male elephants and that the presence of urine dribbling should be the physical signal used for defining this state.     

Influence of age and gender on secretion of anti-Müllerian hormone in Asian (Elephas maximus) and African (Loxodonta africana) elephants

Anti-Müllerian hormone (AMH) secretion was studied in Asian and African elephants varying in age and reproductive status. Overall mean concentrations did not differ (P > 0.05) between species, but were markedly higher in male than female Asian elephants (31.01 ± 4.22 ng/mL and 0.19 ± 0.02 ng/mL, mean ± SEM) and African elephants (40.27 ± 3.18 ng/mL, 0.17 ± 0.04 ng/mL), respectively. Anti-Müllerian hormone secretion was not significantly affected by ovarian cyclicity status (cycling vs noncycling), but was higher (P < 0.05) in prepubertal (0.40 ± 0.10 ng/mL) than reproductive age (8-35 y; 0.18 ± 0.04 ng/mL) and aged (≥ 36 y; 0.16 ± 0.03 ng/mL) females. In males, AMH secretion was not significantly affected by musth status, but was age-related, with higher concentrations (P > 0.05) in prepubertal (49.08 ± 6.11 ng/mL) as compared to aged (≥36 y; 22.27 ± 5.82 ng/mL) bulls; concentrations in mature bulls (8-35 y; 37.01 ± 3.17 ng/mL) were similar to prepubertal and older bulls. We concluded that circulating AMH concentrations in elephants were similar between species and not affected by reproductive status; however, concentrations were significantly higher in males than females, and in younger animals. The diagnostic value of AMH to assess fertility status of individual elephants remains to be determined.     

Rapid chromatography for quantitation of radioimmunoassayable 5alpha-androstane-17beta-ol-3-one and testesterone in ram, bull and boar serum.

A rapid technique for the simultaneous determination of serum 5alpha-androstane-17beta-0l-3-one (DHT) and testosterone (T) was developed and validated. The procedure incorporates a chromatographic step prior to binding with a T-3BSA antiserum. Anakrom columns used to separate DHT from T also isolate other steroids quickly and efficiently. Utilizing this procedure, levels of both androgens have been measured in the intact ram, bull and boar. Concentrations of DHT and T expressed as mean +/- SE in ng/ml were respectively: ram, 0.14 +/- 0.03 and 4.96 +/- 0.80; bull, 0.10 +/- 0.01 and 3.92 +/- 0.80; boar, 0.71 +/- 0.11 and 5.58 +/- 0.64. Whether these minute quantities of serum DHT play important physiological roles, particularly in the boar, is unresolved. Monitoring DHT under varying experimental or pathological conditions may allude to its importance in the domestic male.     

Predictors of testosterone in zoo-managed male African elephants (Loxodonta Africana)

Reproductive complications for both male and female zoo-managed African elephants (Loxodonta africana) contribute to the rapidly declining population. In zoo-managed bull elephants, few studies have explored the potential physiological, physical, social, and environmental factors that influence bull fertility, particularly, androgen production. Testosterone is the essential steroid hormone for male sexual maturation and inadequate concentrations can be detrimental for spermatogenesis. Testosterone, fecal glucocorticoid metabolites, leptin, glucose, insulin, and triglycerides were analyzed from weekly fecal and blood serum samples taken over 6 months from six zoo-managed African elephant bulls (10–19 years of age). Testosterone levels were compared to endocrine factors, weekly social and environmental variables, daily musth signs, and body condition scores (BCS). The glucose-to-insulin ratio (G:I) was the only physiological biomarker found to be positively associated with testosterone. Predictive physical variables included Musth Score (+) and Moderate Exercise (+). Bulls with BCS signifying overweight (BCS 4) had lower testosterone (36.6 ± 1.6 ng/g fecal extraction [FE]) than bulls with healthy BCS 3; 51.2 ± 4.9 ng/g FE). Numerous social variables influenced testosterone concentrations, including Total Contact Day (+), Female Interaction Day (+), Indirect Contact Day (+), Indirect Contact Night (+) and Total No Contact (−). Both percentage of Time Outdoor and Time Mixed positively influenced testosterone, whereas testosterone decreased for percentage of Time Indoors. Each additional daily browse opportunity increased testosterone by approximately 7 ng/g FE. In managed care, the emphasis should be placed on optimizing these predictors of testosterone production to promote bull reproductive health.     

Serum thyroxine and triiodothyronine radioimmunoassay values in the normal ferret

The European ferret, Mustela putorius furo, has become increasingly popular as an animal model in biomedical research. However, certain important normal clinical data have not been established for the ferret. In this study, serum thyroxine (T4) and 3,3',5-triiodothyronine (T3) values were obtained from ferrets by the use of commercial radioimmunoassays. Sera from 44 animals, 31 males (27 intact and 4 castrated) and 13 females (10 intact and 3 spayed) were assayed. Serum T4 values ranged from 1.01-8.29 micrograms/dl for males (mean = 3.24 +/- 1.65 micrograms/dl), and 0.71-3.43 micrograms/dl for females (mean = 1.87 +/- 0.79 micrograms/dl). Serum T4 values of adult female ferrets, juvenile ferrets (less than 1 year old) of either sex, and castrated males were similar to the normal T4 values of the cat, 1.20-3.80 micrograms/dl. Intact adult male ferrets had higher serum T4 values which were more comparable to those of the normal dog 1.52-3.60 micrograms/dl. Serum T3 values ranged from 0.45-0.78 ng/ml for males (mean = 0.58 +/- 0.09 ng/ml), and 0.29-0.73 ng/ml for females (mean = 0.53 +/- 0.13 ng/ml). These values are comparable to those of dogs and cats which are 0.50-1.50 ng/ml.     

Assessment of testicular endocrine function in captive African elephants by measurement of urinary and fecal androgens

Androgen measurements in urine and/or feces represent a potentially important tool for monitoring testicular endocrine function in the African elephant. To assess the feasibility of this approach, the aims of the present study were to: 1) examine the presence and relative abundance of immunoreactive testosterone (iT) and its 5α‐reduced 17‐oxometabolite epiandrosterone (iEA) in African elephant excreta, and 2) compare urine and fecal androgen profiles in animals of different ages and during the musth and non‐musth condition. Urine and fecal samples were collected over periods of up to 3 years from five bulls (ages 7–24 years) living in three mixed social groups. In parallel, indications of musth were recorded by keeper staff as an independent marker of male androgen status. Measurements of iT and iEA were carried out by enzymeimmunoassay (EIA) following methanolic extraction of hydrolyzed urine and lyophilized fecal powder. High‐pressure liquid chromatography (HPLC) of musth phase samples confirmed the presence of substantial quantities of testosterone (T) and epiandrosterone (EA) in both urine and feces. EA was predominant in feces, whereas T was more abundant in urine. In each male, the two androgen measures were significantly correlated (feces, r = 0.71–0.93, P < 0.0001; urine, r = 0.86–0.91, P < 0.0001), as were fecal and urinary concentrations of each of the two androgens measured (r = 0.35–0.77, P < 0.0001). Moreover, in the two oldest males that showed clear signs of musth, levels of iEA and iT were markedly elevated during musth compared to non‐musth periods (differences were significant for feces in both animals, but in urine only for one). Collectively, the data show that measurement of urinary and fecal androgens generates useful information on gonadal status in male African elephants, and as such should provide new opportunities to improve the management and welfare of bulls maintained in captivity, as well as to examine physiological correlates of reproductive function in free‐ranging animals. Zoo Biol 21:27–36, 2002. © 2002 Wiley‐Liss, Inc.     

Electroejaculation, semen characteristics and serum testosterone concentrations of free-ranging african elephants (Loxodonta africana)

A regimented electroejaculation protocol (120 electrical stimulations; 10-30 V) was used to collect semen and characterize ejaculate quality from 9 adult, free-ranging African elephants under anaesthesia. Eight of the 9 ejaculates contained high concentrations of progressively motile spermatozoa. The overall mean ejaculate volume, sperm concentration/ml ejaculate, sperm motility, sperm status and ejaculate pH were 93.3 ml, 2408.6 X 10(6) spermatozoa/ml, 70%, 3.9 and 7.4, respectively. A high percentage (mean 77.5%) of spermatozoa within each ejaculate was morphologically normal. Of the aberrant spermatozoa, 72% had a cytoplasmic droplet defect. When sperm viability was tested in vitro at 37 degrees C, sperm motility rating declined by at least half of the initial assessment within 3.5 h of semen collection. Generally, spermatozoa maintained motility in vitro for less than 6 h. Serum testosterone ranged from 1.4 to 8.2 ng/ml in 4 males evaluated in the morning (07:30-08:00 h). In 4 of the 5 bulls assessed in the afternoon (15:00-18:00 h), testosterone levels were less than 0.9 ng/ml. The remaining bull, evaluated at 16:00 h, had exceptionally high testosterone concentrations (peak 25.6 ng/ml) and a preputial discharge potentially indicative of 'musth'. The present study demonstrates that high quality semen can be collected consistently from the African elephant and that striking differences exist in serum testosterone amongst free-ranging males which may be due, in part, to a diurnal rhythm.     

Frontalin: a chemical message of musth in Asian elephants (Elephas maximus)

Musth is an important male phenomenon affecting many aspects of elephant society including reproduction. During musth, the temporal gland secretions (as well as the urine and breath) of adult male Asian elephants (Elephas maximus) discharge a variety of malodorous compounds together with the bicyclic ketal, frontalin. In contrast, teenage male elephants in musth release a sweet-smelling exudate from their facial temporal gland. We recently demonstrated that the concentration of frontalin becomes increasingly evident as male elephants mature. In the present study, we demonstrate that behaviors exhibited towards frontalin are consistent and dependent on the sex, developmental stage and physiological status of the responding conspecific individual. To examine whether frontalin functions as a chemical signal, perhaps even a pheromone, we bioassayed older and younger adult males, and luteal- and follicular-phase and pregnant females for their chemosensory and behavioral responses to frontalin. Adult males were mostly indifferent to frontalin, whereas subadult males were highly reactive, often exhibiting repulsion or avoidance. Female chemosensory responses to frontalin varied with hormonal state. Females in the luteal phase demonstrated low frequencies of responses, whereas pregnant females responded significantly more frequently, with varied types of responses including those to the palatal pits. Females in the follicular phase were the most responsive and often demonstrated mating-related behaviors subsequent to high chemosensory responses to frontalin. Our evidence strongly suggests that frontalin, a well-studied pheromone in insects, also functions as a pheromone in the Asian elephant: it exhibits all of the determinants that define a pheromone and evidently conveys some of the messages underlying the phenomenon of musth.     

Serum testosterone and dihydrotestosterone changes with age in rat.

Serum testosterone (T) and 5alpha-dihydrotestosterone (DHT) were measured in young, adult and old Albino Wistar male rats using a sensitive and reliable radioimmunoassay, after separating T from DHT by thin layer chromatography. The mean plus or minus S.E.M. for T in young, adult and old rats were 62 plus or minus 11, 250 plus or minus 27 and 125 plus or minus 25 (ng/100 ml) respectively. Serum T was increased in adults (P less than 0.001) and decreased in old rats (P less than 0.01). The mean plus or minus S.E.M. for serum DHT was 8 plus or minus 2, 19 plus or minus 2 and 17 plus or minus 1 (ng/100 ml) for young, adult and old rats respectively. DHT was increased in adults (P less than 0.001), but did not change in old rats.     

Blood prostaglandin A (PGA) levels in normal human subjects

Blood prostaglandin A levels were measured in ten healthy subjects under different conditions of collection and storage.Plasma levels ranged from 1.20 to 1.81 ng/ml (M ± SE = 1.50 ± 0.10) in 5 females to 1.23 to 1.68 ng/ml (1.45 ± 0.09) in 5 males, when centrifuged and frozen immediately after collection. Storate at 4°C for varying times up to 24 hours and at 22°C (room temperature) up to 4 hours did not affect mean plasma concentrations significantly, but increased the range obtained to 1.00 to 2.47 ng/ml for both male and female groups.Serum concentrations differed in males and females and were lower than corresponding mean plasma values for males and higher for females. Mean serum concentrations were 1.77 ± 0.08 ng/ml in females and 1.18 ± 0.05 ng/ml in males and did not change significantly up to 24 hours of storage at 4°C.These results suggest that prostaglandin A assayed in both plasma and serum under the conditions described is stable and should allow for greater flexibility in sampling under different experimental conditions.     

Prostaglandin production by human epidermal cells in vitro: a model for studying pharmacologic inhibition of prostaglandin synthesis

Blood prostaglandin A levels were measured in 10 healthy subjects under different conditions of collection and storage. Plasma levels ranged from 1.20 to 1.81 ng/ml (M + or - SE=1.50 + or - 0.10) in 5 females, to 1.23 to 1.68 ng/ml (1.45 + or - 0.09) in 5 males, when centrifuged and frozen immediately after collection. Storage at 4 degrees Centigrade (C) for varying times up to 24 hours and at 22 degrees C (room temperature) up to 4 hours did not affect mean plasma concentrations significantly, but increased the range obtained to 1.00 to 2.47 ng/ml for both male and female groups. Serum concentrations differed in males and females and were lower than corresponding mean plasma values for males and higher for females. Mean serum concentrations were 1.77 + or - 0.08 ng/ml in females and 1.18 + or - 0.05 ng/ml in males and did not change significantly up to 24 hours of storage at 4 decrees C. These results suggest that prostaglandin A assayed in both plasma and serum under the conditions described is stable and should allow for greater flexibility in sampling under different experimental conditions.     

Treatment of pregnant rhesus macaques with testosterone propionate: observations on its fate in the fetus

Castrated male primates, unlike castrated male rodents, respond to exogenous estrogen by releasing gonadotropin. Although this disparity is unexplained, it may occur because the amount of testicular androgen secreted during a critical period for sexual differentiation is not sufficient to completely androgenize the anlagen of the central nervous system (CNS) in male primates. Therefore, male primates might be incompletely masculinized, and if fetal males were exposed to additional androgen during sexual development, they would no longer display the positive feedback to estrogen that usually characterizes females. Besides the development of mechanisms mediating the release of gonadotropins, questions about relationships between adult male sexual behaviors and the intensity of the androgen stimulus upon developing neural structures are of interest. We tested some of these possibilities by injecting androgen into 8 pregnant rhesus macaques from Days 40 through 50 of gestation, and we compared serum levels of testosterone (T), dihydrotestosterone (DHT), and androstenedione (delta 4) in the fetal circulation with that of 5 untreated control males. Fetal sera (both male and female) from treated pregnancies did not contain significantly greater quantities of T and DHT than sera of intact control males from untreated mothers. The maternal sera, however, contained large amounts of T (125.05 +/- 27.40 [SEM] ng/ml, n = 8) and significant elevations of DHT and delta 4 after treatment. The concentrations of delta 4 in the fetal circulation were significantly elevated (p less than 0.01) in treated fetuses compared to intact control males, probably due to the actions of the 17 beta-hydroxysteroid dehydrogenases in the placenta, the fetus, or both.(ABSTRACT TRUNCATED AT 250 WORDS)     

Endocrine and behavioral changes in male African elephants: linking hormone changes to sexual state and reproductive tactics

Hormones play a crucial role in mediating genetic and environmental effects into morphological and behavioral phenotypes. In systems with alternative reproductive tactics (ART) shifts between tactics are hypothesized to be under proximate hormonal control. Most studies of the underlying endocrine changes behind ART have focused on fish and amphibians rather than mammals and few have investigated the potential interaction between different endocrine axes in regulating shifts between conditional dependent tactics. Using a combination of endocrine and behavioral data from male African elephants we expand on our previously published analysis and show that the initial increase in androgens predates the behavioral shifts associated with reproductively active periods, supporting the role of androgens in activating sexually active periods in males. A strong interactive effect between androgens and glucocorticoids was found to determine the presence or absence of temporal gland secretion and urine dribbling, signals associated with the competitive reproductive tactic of musth, with elevated glucocorticoids levels suppressing the occurrence of musth signals. In addition external environmental conditions affected hormone levels. The presence of receptive females resulted in elevated androgens in dominant musth males but increased glucocorticoids in subordinate non-musth males. The presented data on hormones, behavior and reproductive tactics strongly support an underlying endocrine mechanism for mediating the translation of intrinsic as well as extrinsic local conditions into the conditional dependent reproductive tactics in male elephants via interactions between the hypothalamic-pituitary-gonadal and -adrenal axes.     

Ovarian function in the elephant: luteinizing hormone and progesterone cycles in African and Asian elephants

Serum samples were collected weekly for 3 yr from two female African elephants, for 18 mo from two other female African elephants, and for 2 yr from two female Asian elephants. Animals were not sedated at the time of blood collection. Ovarian cycles, characterized by changes in progesterone and immunoreactive luteinizing hormone (ILH) concentrations, averaged 15.9 +/- 0.6 wk (N = 25) for African females and 14.7 +/- 0.5 wk for Asian females (N = 10). The length of the active luteal phase averaged 10.0 +/- 0.3 wk for African elephants (range 8-14 wk) and 10.6 +/- 0.6 wk for Asian females (range 9-13 wk). Interluteal phases were 5.9 +/- 0.6 wk for African females and 4.2 +/- 0.5 wk for Asian females. One African female (Maliaca) had two extended interluteal phases, both occurring between the months of February and May. Excluding these two periods, there were no differences in the length of the ovarian cycle or the length of the luteal phase between species of elephant. Serum progesterone in both species ranged from less than 50 pg/ml to 933 pg/ml. Average progesterone concentrations during the luteal phase were significantly lower in African elephants compared with Asian elephants (328 +/- 13, N = 30 cycles vs. 456 +/- 23, N = 14 cycles; p less than 0.001). ILH ranged from nondetectable to 11.6 ng/ml. These data suggest that the length of the ovarian cycle in the African elephant is about 16 wk and confirm that the length of the ovarian cycle in the Asian elephant is about 15 wk.     

Sexual differences in 5′-deiodinase activity in the harderian gland of Syrian hamsters and the effect of pinealectomy: Regulation by androgens

Sexual differences on thyroxine 5′-deiodinase (5′-D) in the Harderian gland of Syrian hamsters were investigated. We compared the 24-h profile of 5′-D activity in male and female hamsters, observing a clear rhythm in males but not in females. Female values were always significantly higher than male ones. After pinealectomy day/night variations in male 5′-D activity at the time points studies were abolished, results that are in correlation with serum thyroid hormones. We also studied the regulation by androgen of the enzyme activity. Basal 5′-D activity increased in castrated males and levels fell when animals were implanted with testosterone or its product 5α-dihydrotestosterone (DHT). Female 5′-D activity was also inhibited by androgens. As only the addition of DHT in the presence of epitestosterone, an inhibitor of the conversion of testosterone on DHT, in castrated males was able to decrease 5′-D activity to control animal levels, we suggest a probable direct effect of DHT by itself. © 1996 Wiley-Liss, Inc.     

Urinary androstenedione and luteinizing hormone concentrations during musth in a mature African elephant

Musth has not been well documented in captive African elephants. A 37-year-old African bull elephant in the Kansas City Zoological Park was observed during periods of behavioral musth and non-musth. Androstenedione and luteinizing hormone (LH) concentrations in urine were measured by radioimmunoassay. Urinary androstenedione and LH levels were significantly higher in musth urine than in non-musth samples. A positive correlation (P > 0.001) existed between urinary LH and androstenedione concentrations. These results indicate that musth can occur in a zoo-maintained African elephant and that urinary androgen levels are elevated during musth, possibly as the result of LH stimulation of testicular steroidogenesis.     

Sexual segregation in habitat use by elephants in Chobe National Park, Botswana

We report on a study conducted on free‐ranging African elephants in the woodlands of northern Botswana. We compared bull groups and family units with regard to (1) their patterns of habitat use and (2) their ranging distances from perennial water sources. During the dry season, adult males frequented more habitat types than family units, whereas family units used a wider diversity of habitats than bulls during the wet season. Bulls roamed widely (>10 km) from perennial drinking water in the dry season, when family units congregated within 3.5 km of the rivers. During the wet season, when ephemeral pans were abundant, all elephant groups were found at intermediate distances (5 km) from the rivers. The spacing of elephants in the dry season is consistent with sexual segregation but we reject the hypothesis that this is an outcome of indirect competition for food, because our concurrent studies on elephant feeding ecology found no evidence for intraspecific competition. Instead, we propose that most adult male elephants space themselves to avoid conflict with musth bulls and roam widely in the dry season between discretely distributed feeding ‘hotspots’. The small proportion of males that are in musth remain close to family units to maximize mating opportunities, and family units are unable to range far from water in the dry season. This is due to (1) comparatively high rates of water turn‐over among juveniles and lactating cows and (2) the reduced mobility of neonates.