The movements of southern elephant seal pups in this study fell into three phases. Phase 1, lasting an average of 30 days, consisted of rapid directed dispersal from Macquarie Island with a mean extent of 1169 km. Phase 2, lasting an average of 67 days, consisted of slower and less directed travel. Phase 3, lasting an average of 42 days, consisted of rapid, but less directed than in Phase 1, travel back to Macquarie Island (for six of the seven seals tracked to the end of Phase 3) or to Chatham Island (for seal LM2). We consider it reasonable to assume that Phase 2 primarily represents feeding. This supposition is supported by LeBoeuf et al. (2000), who showed that the proportion of days with reduced travel rate in northern elephant seal tracks was positively related to mass gain between departure and return to land.
The classification of movements into three phases appeared in most cases to be appropriate. However, there were potential difficulties when considering tracks of different durations. For example, in April HM5 increased its travel rate to almost 100 km day−1 for a period of 2 weeks. It then slowed down again and the definition of end of Phase 2 (5-day running mean of daily travel rates last rose above 20 km day−1) was not reached until 9 May. If the SRDL had failed in April the definition of end of Phase 2 would have been satisfied on c. 1 April. Thus, the end of Phase 2 as assigned may have been premature in seals with a shorter tracking duration. In addition, there was variable reduction in travel rate from seal to seal after their initial fast dispersal from Macquarie Island. HM3 decelerated abruptly while LF7 decelerated slowly. Thus, the date of Phase 1–2 transition was sensitive to the defined threshold of 20 km day−1 in LF7, but less so in HM3. In summary, we regard the three phases of movements to be useful and realistic groupings, while at the same time we regard the exact dates and durations with caution.
comparison with adult movements
Hindell et al. (1991) inferred the tracks of 14 southern adult elephant seals from Macquarie Island using sea temperature data and showed that they may forage up to 4000 km from Macquarie Island. The four males and five of the 10 females in that study seemed to travel to the Antarctic continental shelf, while the five remaining females travelled to the Antarctic Polar Front. None of these adult foraging areas matched well with the Phase 2 areas of the pups in this study. However, additional data provided by Slip et al. (1994) show that the area of our Phase 2 south-eastern group is also used by post-moult adult females. Recent tracking studies (SMRU, unpublished) have also suggested that post-moult females sometimes follow similar routes to the south-east group of pup tracks in this study.
influence of intrinsic factors on movement patterns
The nearest neighbour spatial test showed no evidence that sex, year, weaning mass or departure day, affected Phase 1–2 or Phase 2–3 transition locations. Sexual segregation of foraging areas has been demonstrated to varying degrees in both southern (Hindell et al. 1991; Campagna et al. 1995; McConnell & Fedak 1996; Campagna et al. 1999) and northern (Stewart & Delong 1990; LeBoeuf et al. 1993; LeBoeuf et al. 2000) elephant seal adults. Further, LeBoeuf et al. (1996) showed that pups start to exhibit sexual foraging segregation from their third to fourth trip to sea, while Stewart (1997) found segregation was established in males that were 2–4 years old. This segregation has been attributed to the differing energetic requirements of the highly dimorphic sexes in elephant seals (Stewart 1997). While sex may account for a certain degree of variation in choice foraging areas, it appears not to be the whole story. A localized foraging area on the continental shelf off the Falkland Islands used by an adult male from Patagonia (Campagna et al. 1999) was very close to an area used by a female from South Georgia (McConnell et al. 1996). Thus, sexual segregation may be more apparent within rather than between breeding colonies. Our failure, however, to observe sexual segregation in this study does not detract from the hypothesis that any sexual segregation is driven by differing energy requirements since Bell et al. (1997) showed that there was no difference in male and female growth rates of pups returning to Macquarie Island after their first trip to sea. Thus, under this hypothesis, any sexual segregation would not be expected until later in the pups’ development.
influence of environmental factors on movement patterns
The pattern of dispersal from Macquarie Island formed two groupings: a tight south-easterly group and a more diffuse group to the west. Bornemann et al. (2000) also found a close grouping of southern elephant seal pups as they dispersed to the west, avoiding pack and sea ice, from a breeding site at King George Island, off the Antarctic Peninsula. In contrast, most of the adult females in that study were associated with the ice edge. In our study only two pups travelled to the ice edge, and then only for a combined duration of 7 weeks.
The simplest hypothesis to explain the observed south-east pup dispersal from Macquarie Island is that the pups passively drifted in the ocean currents. In the vicinity of Macquarie Island the predominantly easterly Antarctic Circumpolar Current (ACC) is diverted to the south-east by the Campbell Plateau for approximately 700 km. In this region the current can reach speeds of 0·8 m s−1. The path taken by a passively drifting particle at 100 m depth is shown by the 0 m s−1 swimming speed simulated track in Fig. 3d, and it is clearly inconsistent with the seal track data. However, further inspection of the track data suggests that the addition of the simple rule ‘swim magnetic east’ may produce trajectories similar to the tracks of the south-eastern group. The resultant trajectories at a variety of swim speeds show a general similarity to the south-eastern tracks, but not sufficiently so to reject the hypothesis that swimming at a constant magnetic bearing plays no role in navigation. The gross role of ocean currents in dispersal thus remains uncertain in the south-east group, but it certainly appears to have no role in the movements of the western group. In another satellite telemetry study, Loughlin et al. (1999) hypothesized that the movements of male northern fur seals (Callorhinus ursinus L.) in the Bering Sea and North Pacific Ocean were influenced by surface currents. However they concluded that individual seal tracks were, for the most part, independent of surface currents.
The Phase 2 tracks of the south-eastern grouping were within the 2–5 °C sea surface temperature isotherm band. Otherwise, there was no evident correlation with the available remotely sensed sea surface temperature or colour data. However, these data sets were integrated over time and space and may thus have obscured any relationship with smaller scale or transient ocean surface features.
The return route during Phase 3 was generally less directed than the outward Phase 1. The Phase 3 track of seal LM2 was exceptional in that it travelled north in a directed fashion to the Antipodes Islands and then to Chatham Island. Since this seal had never previously been to either of these islands it could not have been making use of a spatial memory map. However, the directed nature of its approach suggests that the seal either sensed the bearing to these islands from many hundreds of kilometres away or that it followed another animal which had been there before.
Seven of the eight seals which were tracked back to, or close to, Macquarie Island approached the island from the south-east, including one animal that had travelled from the western group. This funnelled approach from the south-east was against the prevailing current of the ACC and contrasts with the wide spread of initial departure bearings. Such a pattern of behaviour is consistent with the hypothesis that final navigation back to Macquarie Island involves the detection of a down current, or down wind, chemical signature of Macquarie Island. The use of such olfactory cues has already been suggested in the migration of green turtles (Chelonia mydas L.) to Ascension Island (Koch, Carr & Ehrenfeld 1969) and is well documented in salmon (Oncorhynchus spp.; Dittman & Quinn 1996). The chemical plume hypothesis could be readily tested by using the techniques of Oliver et al. (1998) and translocating pups up and down stream from Macquarie Island.
These south-eastern group Phase 2 tracks were bounded to the south-west by the Pacific Antarctic Ridge. A similar association of elephant seal tracks with subocean ridges and seamounts has previously been observed by a number of workers (McConnell et al. 1996; Jonker & Bester 1998; Bornemann et al. 2000). They suggested that seals were attracted to increased prey density due to enhanced production caused by local upwelling. In this study we suggest that the mechanism connecting the Pacific Antarctic Ridge and foraging in Phase 2 is via the influence of the ridge on the eastwards flow of the Antarctic Circumpolar Current (ACC) (Gordon, Molinelli & Baker 1978). The boundary of the ACC, and the positions of the major frontal systems within it (Orsi et al. 1995) are shown in Figs 1 and 3, and the southern boundary of the south-east group Phase 2 tracks is clearly aligned close to the southern ACC front (SACCF). We should, however, be aware that Orsi mapped the positions of ACC and its frontal systems using historical records, and that the position of the fronts associated with the ACC can vary significantly through time (Moore, Abbott & Richman 1999; Pakhomov, Ansorge & Froneman 2000; Trathan et al. 2000). Thus, there is a degree of uncertainty in the actual position of the ACC fronts and boundaries during our study years. However, Nicol et al. (2000) have recently shown the importance of the area to the south of the Southern Boundary of the ACC (SB-ACC) on local productivity at all trophic levels. Tynan (1998) also documented concentrations of krill (Euphausia superba Dana) and sperm whale (Physeter macrocephalus L.) near the SB-ACC (as mapped by Orsi) and concluded that the SB-ACC ‘provides predictably productive foraging for many species, and is of critical importance to the function of the Southern Ocean ecosystem’. Our study suggests that the area around the southern extent of the ACC is also important as foraging grounds for a significant proportion of southern elephant seal pups.
The pattern of tracks we have observed is not random and is likely to be the result of a combination of intrinsic and extrinsic factors. It is difficult to know if the numerous coincidences of tracks that have occurred are the result of chance encounters or some combination of the use of environmental cues and/or communication links with conspecifics or other animals. However, the frequency of such encounters encourages further observation and offers hope that clues to the cues that animals use to locate foraging locations may be gained from them.
implications for maternal investment
We have made a preliminary attempt to use the behaviour of naïve pups at sea, in conjunction with estimates of the stored resources they have when they depart, to examine the consequences of maternal expenditure on survival to nutritional independence. We emphasize that this is currently based on a number of assumptions that need further study to refine and validate but we believe it remains a useful exercise. We did not measure body composition in our study animals, but assumed that their body composition at weaning and the composition of the mass they lost were similar to that reported for southern elephant seals elsewhere by Carlini et al. (2001). We make the conservative assumption that pups at sea will use materials at the same rate as they do while ashore. While we expect that energy requirements are likely to be higher at sea, it is possible that opportunistic feeding balances this to some degree. And finally, we assume that 10% body fat reserve and 30% protein loss represent critical levels for survival. It is important that all these assumptions are examined and we hope that their use here will draw attention to their importance.
Several of the implications of this exercise deserve special emphasis. Based on the above assumptions, the time taken for animals to enter Phase 2 (the assumed foraging phase) is very close to estimates of critical time for smaller pups whereas larger ones seem to have a greater margin of safety. The calculations also suggest that protein stores as well as blubber stores should be considered when examining the value of maternal expenditure. While animals adapted to prolonged fasting can reduce absolute and relative protein utilization, adequate protein supplies are nevertheless required during fasting (Dulloo & Jacquet 1999). The consequences of the fat–protein balance at weaning also extend to diving ability. In this same set of pups, Hindell et al. (1999) showed that larger pups displayed enhanced diving capabilities, possibly as a result of their greater lean body mass. Thus, the optimal balance of fat and protein resources at weaning may be a complex response to the need to avoid starvation before reaching distant foraging areas and the ability to dive adequately once there. Additional factors such as the role of fat in thermal insulation may also influence this balance. For these reasons and those listed in the paragraph above, detailed studies of body composition in the pup’s first year of life seem essential for understanding maternal expenditure strategies.
The function of the post-weaning fast, common in many species of phocids (Arnbom et al. 1993) is not clear but it seems likely that it has some important physiological function (Blackwell & Leboeuf 1993; Pattersonbuckendahl et al. 1994; Thorson & LeBoeuf 1994; Falabella, Lewis & Campagna 1999) or social (Arnbom et al. 1993; Modig, Engstrom & Arnbom 1997) function because it is always present and is expensive in terms of the amount of body stores consumed to support it. It also acts to isolate pups from their mothers and may deny them the opportunity to capitalize on the mothers foraging experience. Our results suggest that the pup’s decision when to terminate the fast is a particularly critical and complex one. Arnbom et al. (1993) showed that large pups tend to both remain on shore longer and depart heavier, presumably gaining advantage both from the fast and extended survival time at sea before finding food. Animals must leave anticipating a significant further period without food. Physiological signals initiating departure must therefore come in advance of the acute consequences of fasting, in some way taking into account an expectation of the time to successful foraging.
It seems likely that year-to-year variability in oceanographic conditions (Sharhage 1988) will influence the time it takes pups to find food and the rate at which it can be consumed. Thus, we would predict a significant variability in pup survival – body reserves that are adequate in one year may be inadequate or superfluous in others. Indeed, Hindell et al. (1991) have shown that first year survival at Macquarie has ranged from 42–46% to as little as 2% during a period of rapid population decline. In northern elephant seals, LeBoeuf et al. 1994b) found that first year survivorship ranged annually between 20% and 49%. However, the response of pup survivorship to oceanographic conditions will depend upon the extra level of maternal expenditure that is provided to insure pups against the consequences of occasional, extreme years. If mothers routinely provide superfluous resources to their pups, a decrease in survivorship would only be apparent in years when the oceanographic conditions were extremely unfavourable. The pattern of pup survivorship would be further complicated if the mothers’ expenditure was modified by one, or a series of many, unfavourable years. Such a hypothesis is testable with long-term parallel biological and oceanographic studies. Finally, we should be aware that the maternal cost would be reduced if the pup were capable of converting any superfluous resources into increased growth, or capacity for growth or subsequent fitness.
There could be a trade-off between the mother’s material expenditure on her pup and the information the pup has, either from the mother herself or other sources, to help it find food. If food sources were local, reliable and/or the pups were directed to them by information from their parents, then less reserves may be required to give a reasonable expectation of survival until finding food. In such a scenario mothers could give less, and the value of a large expenditure would be reduced. Thus, pups would not require as large a reserve to survive and an extended post-weaning fast would be less costly. However, if food were distant, patchy and unpredictable the value of a given maternal material expenditure may be greater. The availability of good information and cues to direct pups to food could influence the time and effort needed to find food and thus reduce the level of maternal expenditure required.
In summary, maternal expenditure strategy in southern elephant seals would appear to be a complex response to many factors. However, a necessary key to understanding the importance of these factors is an appreciation of the difficulties faced by naïve pups on their road to nutritional independence in a new and variable environment. The elephant seal population on Macquarie Island has declined markedly since the 1950s (Hindell 1991), and is now less than half what it was. In recent years there has been a steady decline of about 1·6% per year (Hindell, Slip & Burton 1994). A number of theories have been put forward to explain why the Macquarie and other southern Indian Ocean population are declining while those of the South Atlantic are stable or increasing (Hindell et al. 1994). We suggest that, whatever the cause for the decline, the complex relationship between maternal expenditure, pup condition at weaning and long-term trends in oceanic conditions needs to be understood before any convincing conclusions can be drawn. This basic biological information can only come from integrated, long-term studies of seal behaviour and oceanography.