Journal of Comparative Physiology A

, Volume 196, Issue 2, pp 147–154

Effects of capture stress on free-ranging, reproductively active male Weddell seals


    • Marine Mammal Research Group, Graduate School of the EnvironmentMacquarie University
  • Emma Turner
    • Marine Mammal Research Group, Graduate School of the EnvironmentMacquarie University
  • Ailsa Hall
    • Sea Mammal Research Unit, Scottish Oceans InstituteSt Andrews University
  • Joseph R. Waas
    • Department of Biological SciencesUniversity of Waikato
  • Mark Hindell
    • Antarctic Wildlife Research Unit, Department of ZoologyUniversity of Tasmania
Original Paper

DOI: 10.1007/s00359-009-0501-0

Cite this article as:
Harcourt, R.G., Turner, E., Hall, A. et al. J Comp Physiol A (2010) 196: 147. doi:10.1007/s00359-009-0501-0


Physiological stress responses to capture may be an indicator of welfare challenges induced by animal handling. Simultaneously, blood chemistry changes induced by stress responses may confound experimental design by interacting with the biological parameters being measured. Cortisol elevation is a common indicator of stress responses in mammals and reproductive condition can profoundly influence endocrine response. We measured changes in blood cortisol and testosterone induced by handling reproductively active male Weddell seals (Leptonychotes weddellii) early and late in the breeding season. Weddell seals have the highest resting cortisol levels of all mammals yet showed a clear, prolonged elevation in cortisol in response to capture. Responses were similar when first caught and when caught a second time, later in the breeding season. Baseline testosterone levels declined over the breeding season but were not altered by capture. Administering a light dose of diazepam significantly ameliorated the cortisol response of handled animals without affecting testosterone levels. This may be an effective way of reducing acute capture stress responses. Male breeding success in years males were handled was no different to the years they were not, despite the acute capture response, suggesting no long-term impact of handling on male reproductive output.


Marine mammalsLeptonychotes weddelliiHandling stressCortisolAntarctica


Wild vertebrates are often temporarily captured, manipulated and then released for the purposes of physiological and behavioural research or for mark-recapture studies to assist the conservation and management of populations (Ortiz and Worthy 2000). The animals may be manually restrained or anaesthetized for varying lengths of time in order to carry out research procedures such as morphometric measurement, attachment of telemetry devices, obtaining tissue samples, or long-term marking of individuals (Ramsey and Stirling 1988; Kelly 1996; Hewison et al. 1999; Powell et al. 2000; McMahon et al. 2005). In most countries, statutory laws designed to protect the health and welfare of the animals being studied regulate manipulations conducted by researchers. It is therefore well recognized that such procedures are potentially stressful to the study animals, where stress is defined as “an abnormal or extreme adjustment in the physiology of an animal to cope with adverse effects of its environment and management” (Fraser et al. 1975). Indeed short-term stress responses, during the capture procedure, may have an impact upon the research questions being addressed. The act of capturing the animal may itself alter some of the parameters, such as clinical blood chemistries, being measured, and so an understanding of the magnitude and duration of the acute stress response is critical to ensure the experimental design can account for such changes (Ninnes et al. 2009). If not, the research outcome itself may be compromised, and measures will need to be devised that alleviate this stress response.

Capture stress may have further deleterious effects on individuals beyond the immediate capture event, with potential longer term consequences for the reproductive success or survivorship of the animals being handled (De Villiers et al. 1997; Omsjoe et al. 2009). Such effects may be of particular significance when the handling involves threatened or endangered species or if the research requires accurate measures of reproductive success. If animals are handled during the breeding season capture stress may have profound influences on overall breeding success within and between seasons. Stressors typically induce increased circulating corticosterone which, during breeding, may inhibit the release of reproductive hormones (Sapolsky 1985; O’Reilly and Wingfield 2001; Wingfield and Sapolsky 2003). This trade-off is a choice between reproductive success and future survivorship. Conversely, there are examples of vertebrates that are resistant to showing environmental and social stress responses during breeding periods (Wingfield and Sapolsky 2003). This may be particularly important in seasonal breeders, such as phocid seals, where the breeding season is so short that acute stress responses may be resisted in favour of reproductive success (Wingfield and Sapolsky 2003).

Many researchers have measured the glucocorticoid response to capture and handling to evaluate how an animal reacts to and copes with a brief, stressful period (Wingfield 1994, for review). The typical response is characterized by an initially low circulating concentration of glucocorticoids that increases significantly within the first 2–5 min (Gardiner and Hall 1997; Le Maho et al. 1992; Vleck et al. 2000). The concentration continues to increase until approximately 30 min, when homeostatic feedback is established and glucocorticoid secretion is reduced. At this point, the concentration reaches a plateau and eventually returns to baseline (Wingfield 1994). How long the glucocorticoids remain elevated may be highly variable and the time taken to return to baseline an indicator of the animals ability to cope with stress (Davis et al. 2008).

Repeated capture of individual seals is a standard method that has been used in numerous studies of seal biology for over 30 years (Fedak and Anderson 1982; Anderson and Fedak 1985, 1987; Boyd and Duck 1991; Arnbom et al. 1997; Pomeroy et al. 1999; Lidgard et al. 2001). If handling significantly adversely impacts individuals this may potentially compromise the specific aims of these studies (Engelhard et al. 2002a, b; McMahon et al. 2005; Wilson and McMahon 2006; McMahon et al. 2008). Therefore, it is important to assess the severity of any stressor and how long any effects persist. As part of a larger programme on Weddell seal (Leptonychotes weddellii) ecology, we have undertaken multiple captures of adults during the breeding season (Harcourt et al. 1998, 2000; Hindell et al. 2002; Wheatley et al. 2006, 2007, 2008a, b), and in particular assessed how individual traits influence male reproductive success (Harcourt et al. 2007, 2008). The aims of this study were therefore to assess the effect of capture, sedation and handling on cortisol secretion in reproductively active male Weddell seals. During the handling phase, animals were weighed and measured, telemetry instruments were attached or removed and serial blood samples were collected throughout. Cortisol and testosterone concentrations were measured in each of these blood samples. We then determined whether use of a mild sedative ameliorated the capture stress response. We also assessed the effect of reproductive state as determined by the timing of handling during the breeding season and by circulating testosterone levels, on capture stress. Finally, we assessed whether the extent of handling influenced reproductive output, as measured by the number of pups sired by the males during the season they were handled.


Study site and species

The study was conducted between October 25 and December 14 1999 at Turtle Rock, McMurdo Sound (77.727S, 166.85E), a site that supported a Weddell seal colony of approximately 25 males and 45 females (Harcourt et al. 2000, 2007, 2008). Weddell seals breed on the fast-ice surrounding the shores of Antarctica. Females give birth over a six-week period from mid-October to mid-November, and are thought to mate at the end of the lactation period, approximately 35–45 days post-partum (Stirling 1969; Hill 1987). Therefore, the study spanned the period when males set up and defend underwater territories and females give birth to their pups, through to the period when mating occurs and females are beginning to disperse from the breeding colony (Kaufman et al. 1975). Some males spend the majority of their time under the ice during this period defending territories, only occasionally hauling out to rest for several hours while other males may spend most of their time hauled out, venturing under the ice for only short periods (Bartsh et al. 1992; Harcourt et al. 2008). Territory defence incurs an energetic cost as many males fast while defending territories, although some males appear to offset this cost by supplementary feeding (Harcourt et al. 2008). The cost of fasting is high with territorial males losing on average 2.1 kg/day (Harcourt et al. 2008) but some losing up to 4.1 kg/day (i.e. >1% of body mass per day) over the course of their tenure, a period which may exceed 40 days (Bartsh et al. 1992; Harcourt et al. 1998, 2008). Fatalities, presumably as a result of this energetic cost, can occur during the breeding season (Hill 1987; Bartsh et al. 1992). This cost is also exacerbated by severe agonistic interactions. Males are frequently found with bites to the head, penile and tail regions. By the end of the season, some males show signs of severe exhaustion and barely respond to approaching humans (pers. obs.).

Experimental design

Weddell seals in Erebus Bay, McMurdo Sound have been the subject of a long-term tagging programme since 1973 (Siniff et al. 1977; Testa and Siniff 1987; Cameron and Siniff 2004). As part of this programme, all pups sighted are tagged within 3 days of birth using plastic cattle-tags in the rear flippers. Old and worn-out tags are replaced when observed and now the proportion of the population that is of known age is more than 60% (Cameron and Siniff 2004). As a consequence, all animals in this study could be identified individually. Eleven adult male Weddell seals were randomly selected for capture early in the breeding season (Capture 1, between 3 and 10 Nov) and subjected to the Early Capture Protocol (see below). Animal handling procedures followed standard protocols as recommended by the Scientific Committee for Antarctic Research Specialist Group on Seals (Erickson and Bester 1993) and as approved by the University of Waikato Animal Ethics Committee. These same 11 animals were recaptured at a later date in the breeding season (Capture 2 between 23 November and 9 December). A second group of animals were also captured (n = 6) during the second capture period to assess the effect of previous handling on capture stress and to determine the effect of changes in reproductive status that may have occurred over the course of the breeding season (see below). Time of day was recorded at capture to account for any diurnal changes in cortisol, although Barrell and Montgomery (1989) reported no diurnal changes in cortisol levels in Weddell seals during periods of continuous daylight such as the Austral spring in this region.

Early capture protocol

The capture of male Weddell seals was restricted to clear and calm weather conditions, with no captures were attempted if wind speed exceeded 25 km h and/or ambient temperature dropped below −20°C.
  1. 1.

    Seals were approached and physically immobilized with a head bag (Stirling 1966).

  2. 2.

    At approximately 5 min (range 1–9 min), a 20-ml blood sample was drawn into heparinzed Vacutainer tubes (Beckton-Dickenson, North Ryde, NSW, Australia 2113) from the femoral vein at the base of the rear flippers using a 1½ inch, 18-g needle, or epidurally using a 5½ inch 14 g needle (baseline sample).

  3. 3.

    The animals were restrained within a 3 m pole net and weighed by hoisting them with a chain block suspended from a tripod using a load cell (GEC Avery, New Zealand, capacity 2,000 kg, accuracy ± 1 kg).

  4. 4.

    Following weighing, approximately 35 min after first capture, animals were sedated with an intramuscular injection of 2.0 mg kg−1 dose of ketamine (Ketamil injection—ketamine as hydrochloride) and 0.01 mg kg−1 of diazepam. The head bag remained in place until the sedative took effect.

  5. 5.

    At 50 min (range 41–78 min) post capture, a second blood sample (10 ml) was taken.

  6. 6.

    The dorsal arch of the animal’s back was dried and cleaned by rubbing with cloths and 100% alcohol. Between one and three telemetry devices were glued to the fur using Araldite # 2017© epoxy. The head bag was kept in place throughout the gluing.

  7. 7.

    Once the epoxy had set and the devices were firmly affixed, the head bag was removed and the research team stepped away from the animal and monitored it from approximately 10 m, but preventing it from entering the water. Although not physically restrained the animal was not free to return to the water until the protocol was complete at 180 min (as below).

  8. 8.

    At 120 min (range 117–130 min) post capture the animal was approached and head-bagged for a third 10-ml blood sample. The animal was monitored as described above.

  9. 9.

    At 180 min (range 181–238 min) post capture the animal was approached and head-bagged for a final 20-ml blood sample, after which the animal was released and left alone.


The time each blood sample was taken was recorded. Blood samples taken at the 5 min mark post-capture were categorized as ‘baseline’ samples. These samples do not represent basal cortisol levels (i.e. levels unaffected by the capture process) but act instead as an interval by which to compare samples taken later in the capture period.

Variations in the standard protocol

Late recapture (Capture 2)

Recaptured males were sampled as for the early capture with two changes (1) telemetry devices were removed by cutting the fur under the device with a scalpel. (2) Five of the 11 males were given 0.4 ml of diazepam (10 mg/2 ml) intramuscularly. No ketamine was administered and the remaining six individuals were not sedated.

Late capture

Six additional male seals were captured at Turtle Rock during the same period as the re-captured animals. They were sampled as for the early capture, except that these animals were not sedated. However, their dorsal surface was rubbed for approximately 30 min to simulate device attachment or removal.

Blood treatment and storage

Blood samples were taken to a nearby heated hut within 90 min of extraction, where the red cells were allowed to settle at room temperature for approximately 3 h and then centrifuged at 2,500 rpm for 9 min. The plasma was extracted and frozen at −20°C in a portable freezer until they were transported to Australia on dry-ice. All samples were frozen within 8 h of initial blood extraction.


Cortisol and testosterone concentrations were measured in the plasma samples, using hormone-specific Spectra radioimmunoassay (RIA) kits (Australian Pty Ltd, Sydney). For cortisol 20 μl of plasma (1:4 dilution with TRIS Buffer) was added to duplicate assay tubes pre-coated with an anti-human cortisol antibody. The samples and a set of serially diluted standards (0–2,000 nmol l−1) were incubated with 0.5 ml of 125I cortisol in a 37°C water bath for 2 h. The liquid was decanted, the tubes rinsed with 1.0 ml of distilled water, allowed to dry and counted for 1 min at approximately 20,000 cpm in a Gamma counter. Concentrations were calculated from the standard curve. Samples of pooled plasma were assayed to determine the intra-assay and inter-assay coefficients of variation (CVs). Intra-assay CV ranged from 4.1 to 2.9% and inter-assay CV from 12.2 to 9.9%. Recovery rates were calculated using pooled plasma added to 5, 20 and 50 μl of the 500 nmol l−1 cortisol standard. Mean recovery was 81.3%. The sensitivity of the assay, defined as the detectable concentration equivalent to twice the standard deviation of the zero-binding value (B0), was <12 nmol l−1.

The testosterone radioimmunoassay was the same format as the cortisol assay, using 25 μl of undiluted plasma. The serially diluted standards ranged from 0 to 50 nmol l−1) The intra-assay CV was between 5.3 and 3.7% and inter-assay CV between 14.0 and 11.1%. Mean recovery was 84.7%. The sensitivity of the assay was <0.2 nmol l−1.

Handling and reproductive success

In a parallel study, the mating success of all males sighted at Turtle Rock during the breeding seasons of 1997–1999 was measured using paternity analysis (Harcourt et al. 2007, 2008). Mating success could therefore be measured for males both in years they were intensively handled and for years in which they were not handled at all.

Data analysis

We used a linear mixed effects model in SPlus (TIBCO Software Inc, California) with individual as a random effect and time, season and sedative as fixed effects. We investigated changes in cortisol levels during handling (TIME 5, 50, 120 and 180 min post capture), the influence of season (DAY early vs. late) and whether the use of a sedative mediated any cortisol elevation at instrument removal (SEDATIVE). As testosterone levels may vary as the breeding season progresses, initial testosterone levels at each capture event were included as a covariate in the model to account for potential changes across the breeding season. The effect of handling on mating success was assessed using animals as their own controls with a paired t-test comparing each animal’s success in years that they were handled with years when they were not.


Cortisol response during capture

Changes in cortisol plasma levels during the capture period (5, 50, 120 and 180-min post capture) for the 11 adult male Weddell seals captured early in the breeding season and re-captured later in the breeding season are shown in Fig. 1. All individuals showed a significant increase in circulating cortisol during capture with levels increasing from 1,115.72 ± 66.47 nmol l−1 at 5 min to 1,225.64 ± 63.34 nmol l−1 at 50 min with a peak near 120 min (1,345.74 ± 73.49 nmol l−1) (Table 1). Levels remained elevated at 180-min post-capture (1,365.15 ± 109.64 nmol l−1) (Fig. 1, Table 1). The response was similar for animals at their first capture early in the season and when they were re-captured late in the season (Linear mixed effects model, Table 1, Fig. 1).
Fig. 1

Cortisol levels rise significantly over the course of a capture episode both at first capture early in the season (n = 11) and when animals are recaptured later in the season (no sedative animals, n = 5)

Table 1

ANOVA table from linear mixed effects model (with individual as random effect, all other variables fixed effects) examining changes in cortisol levels in relation to: the use of a sedative during capture (Sedative); when during the breeding season the capture occurred (Day); and the time during the capture event that the sample was taken (Time)




F value

p value































The model shows (1) that cortisol levels were significantly elevated over the course of capture (Time) and (2) that cortisol levels were more elevated during a capture event for animals that had not received a sedative at the second capture

Effects of season and previous handling on cortisol and testosterone levels

There was no significant difference in the initial cortisol levels between the early and late season recapture of the same individuals, and animals caught for the first time late in the season showed similar responses to those being re-captured at the same time. Although testosterone levels at capture were significantly higher early in the breeding season (early = 14.31 ± 2.59, late = 5.63 ± 1.47 nmol/ml) there was no effect of capture and handling on circulating testosterone (Table 2). The testosterone concentrations did not significantly change over the course of a capture event either early or late in the breeding season (Table 2, time effect p = 0.438). When testosterone levels were included as a covariate in the mixed effects model, there was no effect on the cortisol—time relationship (F = 0.3484; p = 0.5569), indicating that cortisol elevation during capture was independent of seasonal changes in testosterone, despite a high level of variation due to differences in breeding status.
Table 2

ANOVA table from linear mixed effects model (with individual as random effect, all other variables fixed effects) examining changes in testosterone levels in relation to the use of a sedative during capture (Sedative); when during the breeding season the capture occurred (Day) and the time during the capture event that the sample was taken (Time)




F value

p value





















The model shows that (1) testosterone was significantly lower in the latter part of the breeding season (DAY) but (2) was unaffected by whether or not the animal had been sedated (sedative) and (3) did not change over the course of individual capture events (Time)

Effects of sedation

For recaptured animals cortisol levels were significantly elevated when handling involved no administration of anaesthetic agents (Table 1), but for those given 0.4-ml diazepam (equivalent to 0.0053–0.0069 mg/kg) cortisol levels were not elevated at any time during the capture event and were over 300 nmol l−1 lower at 120 min (1503.45 ± 121.48 vs. 1088.3 ± 75.70 nmol l−1; Fig. 2). In contrast to cortisol, testosterone levels throughout capture events were unaffected by the administration of diazepam (Table 2, sedative effect p = 0.227).
Fig. 2

For animals being recaptured, a low dose of sedative (Diazepam 0.4 ml at a concentration of 10 mg/2 ml) delivered intra-muscularly mitigates the extent of the stress response to handling

Handling and reproductive success

In addition to the 17 animals included in this study, a further nine males were handled in the three breeding seasons 1997–1999. The mean breeding success of males for the year they were handled (determined by the number of pups they had fathered that were born the following year) was 0.56 ± SEM 0.22, and for the years in which they were not handled the mean breeding success was 0.46 ± 0.13 indicating that there was no difference in reproductive output that could be attributable to capture handling (t = 0.365, df = 25, p = 0.72).


Weddell seals have been reported as having the highest resting cortisol levels of any mammalian species and the high levels for the adult males in this study (930–1,528 nmol l−1) are consistent with those earlier reports (Liggins et al. 1979; Barrell and Montgomery 1989; Bartsh et al. 1992; Liggins et al. 1993). The very high cortisol levels found in Weddell seals are comparable to those found in other Antarctic phocids such as crabeater seals Lobodon carcinophagus and leopard seals Hydrurga leptonyx (Liggins et al. 1993).

Despite the very high initial cortisol levels, male Weddell seals that were captured and handled still demonstrated a significant increase in circulating cortisol over the course of the three hour period of restraint. The time course of this stress response was similar to adrenocortical responses to capture and handling reported in a number of other studies of seals, albeit starting at a higher initial level (Thomson and Geraci 1986; Gardiner and Hall 1997; Engelhard et al. 2002a, b). Seals produced this elevation in cortisol during capture whether sedated by immobilizing agents (Ketamine/Diazepam) or physically restrained without chemical induction (Fig. 1). This contrasts with a study by Engelhard et al. (2002a) who found that while physically restrained Southern elephant seal (Mirounga leonina) pups at 11 and 21 days of age produced a strong cortisol response due to handling, their chemically immobilized mothers produced only a mild cortisol response. Engelhard et al. (2002a) suggest that the mild responses of elephant seal mothers do not necessarily imply that the handling event itself was only a mild stressor, but rather that a variety of physiological responses to stress are suppressed due to lactation. Male Weddell seals are also likely to be chronically stressed during the breeding season as they fast and fight fiercely for access to females (Bartsh et al. 1992; Harcourt et al. 2008). Fasting males may lose as much as 25% of their initial body mass by the end of the breeding season (Harcourt et al. 2008). Simultaneously testosterone levels decline (this study). Despite the handling events and significant hormonal changes there was no relationship between the stage of the breeding season the animals were captured and the magnitude or duration of the capture stress response. There was also no effect of prior handling experience as naïve animals late in the breeding season showed an almost identical response to those that had been handled previously. This again contrasts with the results of Engelhard et al. (2002a) who showed that repeated handling (3–4 events) of female elephant seals resulted in chronic dampening of adrenocortical responsiveness. However, in this study we only handled animals twice in a season and this corresponds to the moderate treatment group in the study by Engelhard et al. (2002a). The moderate treatment group in that study also demonstrated little response to being handled twice. Hence handling may be regarded as a short term stressful event but two events are not sufficient to induce habituation nor, conversely to increase the stress on subsequent capture.

Serial weighing to assess mass and body composition changes is widely used in studies of reproductive success in large mammals (Fedak and Anderson 1982; Anderson and Fedak 1987; Boyd and Duck 1991; Arnbom et al. 1997; Lidgard et al. 2001; Harcourt et al. 2008). In this study, we assessed the mating success of males in years in which they were serially weighed when compared with years when they were not handled. There was no difference in the mating success of these individuals. This further suggests that although the handling event may induce an acute short-term stress response, two captures during a single breeding season does not adversely affect their subsequent behaviour and critically their reproductive success. This concords with findings from recent studies that have shown that while handling induces a short-term stress response, it does not affect either weaning mass (Engelhard et al. 2001) nor first year survival in elephant seals (McMahon et al. 2005), nor the likelihood of calving in reindeer (Omsjoe et al. 2009).

Sedation reduces/mitigates extent of response to handling

Weddell seals are well known to be behaviourally passive in the face of disturbance. They do not flee when approached by over-snow vehicles or people on foot (Kaufman et al. 1975; van Polanen Petel et al. 2007). However, the elevation of cortisol levels during a capture event suggests that like other wildlife, handling is a stressful event for these animals even in the absence of an anticipatory response. Morton et al. (1995) assessed plasma cortisol concentrations in 712 individual animals of 18 wildlife species after either physical or chemical restraint, tranquillisation or trauma. The authors determined that cortisol levels appeared to rise after capture in all the species examined except the Cape Buffalo (Syncerus caffer). They too found that plasma cortisol levels were less elevated in animals handled using sedation than those physically restrained, suggesting that the former is less stressful albeit potentially more dangerous for the animal. Anaesthetized animals had significantly lower cortisol levels than physically restrained animals, with the highest levels being reported among those that suffered trauma or subsequently died during the handling event (Morton et al. 1995). In this study, Weddell seal males immobilized with ketamine/diazepam or physically restrained showed similar cortisol stress responses. However, in the latter part of the season there was complete amelioration of the cortisol stress response in animals given a low dose of a relatively mild and safe sedative. This suggests that the administration of mild sedation even in the absence of a behavioural response may be an efficacious and cost effective way of reducing stress in wild seals subject to any repetitive handling procedure.


We would like to thank the staff of Scott Base and Antarctica New Zealand who provided excellent field support for 4 years of this study. We thank Dudley Bell, Tony Dorr, and Sarah Winter for field assistance. Randy Davis, Terrie Williams, Tom Gelatt and Mike Cameron provided invaluable support at various times. Two anonymous reviewers provided comments which improved the manuscript. The study was supported by the Seaworld Research and Rescue Foundation, the Australian Research Council, the Graduate School of the Environment, Macquarie University, the Department of Biological Sciences, University of Waikato and the Antarctic Scientific Advisory Committee. Permission to conduct the study was obtained from the Environmental Assessment and Review Panel of Antarctica New Zealand, the Department of Conservation, New Zealand and the Animal Ethics Committee of the University of Waikato.

Copyright information

© Springer-Verlag 2010