Twenty—
Expenditure, Investment, and Acquisition of Energy in Southern Elephant Seals

Michael A. Fedak, Tom A. Arnbom, B. J. McConnell, C. Chambers, Ian L. Boyd, J. Harwood, and T. S. McCann

ABSTRACT. Information on the expenditure and investment of energy in southern elephant seals, Mirounga leonina , was collected during breeding and molt over four field seasons at South Georgia. Weight and body composition changes of mothers, pups, and breeding males were monitored during the breeding season. These changes were also measured in adult females, before and after the 70-day period when animals fed at sea between breeding and molt. During this period, information on foraging movements and behavior was gathered using purpose-built satellite-relay data loggers. Body composition changes were measured using isotope dilution techniques.

Breeding energetics information is discussed in relation to the evidence for differential investment in male and female pups. Large females produce larger pups, both at birth and weaning. Male pups are born larger than female pups. However, there is no evidence that mothers invest more energy (either relative or absolute) in male pups after birth once female size and birth weight are taken into account.

Foraging movements and diving behavior are discussed in terms of the oceanography of the foraging area and possible constraints placed by prey consumption on the seals' dive behavior. We suggest that the long distance travel of females to distant feeding locations may be advantageous in providing for the requirements for reliable food sources in a long-lived, uniparous mammal. Dive characteristics changed during the different phases of activity in foraging animals in relation to the average daily velocity of the animal, water depth, and undersea topography.

Southern elephant seals divide their year between land and sea. While on land, they breed and molt, expending energy and material that was stored in their bodies while foraging at sea over the remainder of the year. Some energy is invested in the production of young and some in new skin and hair, while the remainder is metabolized to support the animal during the fasts associated with these activities. Southern elephant seals separate these

periods of net energy loss, both geographically and temporally, from their foraging efforts. This presumably allows them some freedom in their choice of the location of their breeding and molting areas and can help to insulate them from local changes in the abundance of prey. The sharp distinction between periods of net energy gain and loss and the temporal separation of breeding and molt make it possible to study each of these phases of the life history separately. Because there is no feeding during either the breeding or molting periods and the animals are on land and accessible during these times, the behavior and energy expenditures of these activities can be studied using relatively straightforward techniques.

The same cannot be said of studies of the periods of energy gain while the animals are at sea. Because the southern elephant seals that breed on South Georgia have a very large number of widely scattered breeding and molting sites from which to choose and because travel on the island (and to other islands) is difficult, recovery of time-depth recorders from animals is uncertain; this makes the use of telemetry advantageous. However, the scale and remoteness of the Southern Ocean make this difficult. The study of the dive behavior of the species therefore has lagged behind that of northern elephant seals (M. angustirostris ). However, a combined data logger/Argos transmitter has been developed by the Sea Mammal Research Unit (SMRU) to gather detailed information about their movements and behavior while at sea. These devices are now producing a flood of new information on the southern species.

Here, we bring together information collected as part of a joint program involving the SMRU, the British Antarctic Survey, and the University of Stockholm on (1) breeding energetics, (2) parental investment, and (3) energetics and behavior of molting and foraging in this species. The information was collected over four field seasons from 1986 to 1991 at Husvik, South Georgia. It presents work in the process of analysis and publication and is preliminary attempt to synthesize the energetics of the life history of this species.

Using a combination of techniques (serial weight changes and isotope dilution measurements of body composition), we provide information on the energy expenditure of female southern elephant seals during breeding and molt and relate this to their size, their energy stores, and the sex and growth of pups. We use this information to consider the evidence for differential parental investment in this species. We briefly consider the reproductive effort of harem males and compare this with that of females. We then present information on foraging behavior of females and their weight gain during the period at sea between breeding and molt using serial weight changes and position and dive depth/velocity information provided by the satellite-relayed data logger.

Expenditure and Investment While Ashore

Overview of Reproductive Season and Annual Cycle

Southern elephant seals have a circumpolar distribution in the Southern Ocean. They breed during October and November on a small number of subantarctic islands and mainland sites in South America (Ling and Bryden 1981; Laws 1984). Approximately half of the world population breeds at South Georgia (54°S, 35°W) (McCann and Rothery 1988). Female southern elephant seals may begin to breed at 3 years of age, but the majority do not come ashore to breed until age 4 or 5 (Laws 1960; McCann 1980, 1981). Pregnant female southern elephant seals begin to arrive on beaches on South Georgia in mid-September, after some males have made an appearance. New females continue to appear until early November. They give birth about one week after arrival and nurse their pups for 18 to 23 days. Mating occurs after an average of about 22 days. Pups are weaned when females leave the beach but then remain ashore for 3 to 6 weeks. After breeding, females spend around 70 days at sea, then come ashore again for approximately one month to molt (Laws 1956). Males spend more time ashore while breeding and molt about one month later than females.

An Important Proviso

Breeding female elephant seals vary greatly in size: the largest females may weigh three times more than the smallest. We have tried to include in the study significant numbers of the smallest and largest animals on the beaches. Very large and small animals are probably somewhat overrepresented in the sample, and, therefore, the distributions about the means for some of the variables of interest are probably not representative of the population norms. Rather, the sample emphasizes the potential range of values the variables can take and the relationships possible over the size range of females in the population.

Growth of Pups

Table 20.1 summarizes information on weight changes for a sample of females and pups taken in 1986 and 1988 at Husvik, South Georgia. Pups are born weighing an average of 43 kg; males average 6 kg heavier than females (McCann, Fedak, and Harwood 1989). Overall, birth weight is positively correlated with the maternal weight, although this correlation is significant in female pups but not in males when the sexes are treated separately. Growth during lactation follows a roughly sigmoid trajectory (fig. 20.1). It may begin during the first day after birth, but in some pups, it may be negligible or even negative for the first 1 to 10 days. A period of constant growth rate follows this postpartum lag. This is in turn followed in

Fig. 20.1
Weight changes in male (arrows) and female (circles) pups. Growth follows a roughly sigmoid
trajectory with a variable lag and some indication of a growth slowdown near weaning.

many cases by a decreasing rate just prior to weaning. Size at weaning is not highly correlated with weight at birth, and the size advantage held by males at birth is often lost by the time of weaning. Average weight at weaning of male and female pups is not significantly different if the size of their mothers and the size of pups at birth are taken into account. This results from the fact that larger females produce larger pups and that the average weight of mothers of male pups was significantly larger than that of female pups, primarily because few small females in the sample had male pups (see discussion below).

Weight Loss of Females

The rate at which females lose weight during lactation depends, in part, on their size; large females lose weight more rapidly than small ones (fig. 20.2a), but they also tend to produce larger pups (fig. 20.2b). Overall, there is no clear difference in the growth of male and female pups (table 20.1) or the weight loss of their mothers when one accounts for differences in the size of the mothers and the birth weight of their pups. However, some small females that produced male pups lose more weight than might be expected, and their male pups seem to grow somewhat faster than average for pups of females of this size (fig. 20.2b). However, there are few small females with male pups in the sample (in spite of the effort to sample small females), and this trend would require much larger sample sizes to establish.

Changes in Body Composition

After weaning their pups, some females leave the beach looking extremely thin, while others (often the largest) look quite fat. This suggested to us that in spite of the fact that larger females produce larger pups at both birth and weaning, there might still be a differential investment with respect to the size or age of the mother. There might be large differences in resources available in small and large females, and although larger females tend to produce larger pups, the amount invested relative to the amount available (relative investment) might vary as a function of size. To measure this investment in terms of energy and materials and to compare it to the resources that mothers of different sizes had available, the body compositions of 45 females were measured at the beginning and end of lactation during 1986 and 1988 (Fedak et al. 1989) using isotope dilution techniques (Reilly and Fedak 1990).

Table 20.2 gives the average use of energy, fat, and protein for the nursing females in the sample, subject to the proviso noted above. On average, approximately 40% of the energy, 47% of the fat, and 17% of the protein in the body is utilized in producing a pup. However, these amounts vary

Fig. 20.2
(A)  The relationship between initial female mass and daily weight loss
split by male (arrows) and female (circles) pups. Large females lose
weight more rapidly than small females; some small mothers of male
pups lose weight rapidly. (B)  The relationship between initial female
mass and male (arrows) and female (circles) pup mass at weaning.
Large females produced larger pups. The relation is significant (p < .01)
for female pups but not for males when sexes are considered
separately. See fig. 20.3 for explanation of symbol sizes.

widely among individual females (fig. 20.3). Some very large females used as little as 30% or less of the total energy or fat available. In contrast, some of the smallest females that produced large pups used almost 60% of the total energy in their bodies and up to almost 70% of the fat. Some of the smallest females used little of their reserves, but these were mothers who produced pups in the lowest quartile of weaning weight.

One way of viewing this distribution of data is to consider females in order of decreasing size. The largest females can produce even the largest pups with small relative investment; indeed, in theory, some have sufficient reserves to produce two. Some midsize females can produce large pups with no more than average relative investment. But small females face a problem: if they produce pups of average size or greater, they will use a very large fraction of their reserves, possibly with deleterious effects on their subsequent reproduction or survival (Huber 1987; Reiter and Le Boeuf 1991). If they produce small pups, these pups may have a reduced probability of survival to adulthood. The best option for small females might be to abort their pup as early in gestation as possible if their own size at parturition will fall below that necessary to produce a viable pup.

Thus, we could view the distribution of investments in figure 20.3 as reflecting a solution to this dilemma. As female size decreases, relative investment increases up to a point; at weights below the median of female size, females produce either normal pups at very high relative investment or very small pups, well below median size. Indeed, the small number of small females with pups of any size in the sample may well suggest that many small females either do not become pregnant or abort before term.

Fig. 20.3
The relationship between initial female mass and the percentage of energy
stores used in nursing male (arrows) and female (circles) pups. Symbol
size reflects pup size in quartiles: the largest symbol indicates pups in the
highest quartile of weaning weight, and so forth. Note the marked
variation in relative investment with changes in female size.

Energetics of Reproduction in Males

The information presented below on weight changes of males is an interim and incomplete presentation of our data. We include it here to complement the data on northern elephant seal bulls (Deutsch et al., this volume).

Male elephant seals on South Georgia may become sexually mature at 5 to 6 years of age, although most harem bulls are probably 9 to 12 years old (McCann 1981). As is true among females, a very wide size range of animals is present on the breeding beaches. Bulls occupying positions within harems typically weigh 1,500 to 3,000 kg, and weights up to 3,700 kg have been reported (Ling and Bryden 1981). Bulls peripheral to harems may weigh as little as 1,000 kg. The duration of the bulls' stay in and around harems is also very variable. In our sample, the longest stay was 73 days; the average was 58 days. Males lost weight at 12 to 15 kg per day while ashore, although one male lost weight at a prodigious 20 kg/day. Typical males lose around 700 kg during their breeding fast, and some lose up to 1,000 kg.

When compared to lactating females, the rate of weight loss and total

weight loss are much greater and more variable. Females lose around 8 kg/day for 23 days or about 180 kg. Much of this loss is material, largely fat and protein, transferred to their pup. Male losses are largely metabolic, probably mostly fat consumed to provide energy for breeding activity. While males are 2 to 8 or more times the size of breeding females, their total losses are only about 4 times as great, even though they may spend double the time breeding and fasting.

These rates of weight loss are larger than similar estimates for northern elephant seal bulls (Deutsch et al., this volume). This is not unexpected, since southern elephant seal bulls are somewhat larger. The spatial organization of harems on South Georgia is also quite different from that on most northern elephant seal beaches. Harems on South Georgia tend to be linearly arrayed along the shoreline and are often separated by unoccupied beach or rocky headlands. This distribution may require a swim if males from one harem are to visit another. The number of interactions between animals could therefore be quite different when compared to northern elephant seals among which harems often spread back from the beach and are more crowded. This might well influence the energy used during breeding.

Behavior and Energetics of the Molt in Females

Less is known about the behavior of animals during the annual molt or the energetic and material requirements of this phase of the life cycle of southern elephant seals. However, we have begun work to find out how much energy animals use during the molt and how much energy they gain while at sea between breeding and molt (Boyd, Arnbom, and Fedak, 1993). During the breeding season of 1990–1991 at Husvik, we captured 20 females late in lactation, weighed them, determined body composition isotopically, and marked or radio tagged them. Some were recaptured when they made landfall at the start of the molt and again 2 to 4 weeks into the molt and the procedures repeated. During the molting period, individual animals were located every few days until they left the Stromness Bay area. Reports of this study are now in press (ibid.), and we can now make some statements about the behavior and weight changes during the molt and about the weight gains while the animals are at sea between breeding and molt.

Adult females begin to arrive in significant numbers at molting sites around Husvik in mid-January (Laws 1960). Roughly one-third of the females tagged at breeding returned to the Stromness Bay area to molt. One female molted in another bay 10 km away in an area inaccessible to us, but we could monitor her presence via a VHF radio. At least two of the radio-tagged seals are known to have molted away from South Georgia on islands near the Antarctic Peninsula.

Several phases seem to occur during the molt in terms of both the mor-

phological changes (Worthy et al. 1992) and, from our observations, the behavior of the animals. When radio-tagged animals first come ashore, they may make landfall at one site and then reenter the water and move to other sites, sometimes many kilometers away. They are often quite mobile for several days before settling down in a "wallow" or other place not immediately adjacent to the water's edge. Once in such a spot, they often remain there for 7 to 14 days while their old hair and skin loosens and falls away. Once this stage is complete and new hair begins to be detectable by touch, they often move out of the wallow to areas where their skin remains relatively clean and dry. During this period, hair growth continues and animals lose the velvet feel and appearance they have when leaving the wallows. They often move close to shore during this time and, on warm days, enter the water occasionally. Once at the shore, they may later move along it to more exposed locations on points and rocky headlands. Some animals were found along the shore for many weeks after leaving wallows and after hair seemed completely regrown. These changes in behavior may be related to underlying thermoregulatory or mechanical needs associated with shedding and regrowth of hair and skin.

During this period, animals lost 4 to 5 kg/day, a rate slightly greater than that of northern elephant seals (Worthy et al. 1992), but the animals in that study were somewhat smaller. This loss is roughly half that during lactation. If animals remain ashore for 30 days, the total loss may approach two-thirds to three-fourths of the lactational loss. The chemical composition or the weight lost is slightly different from that lost during lactation, with 39% of the loss being fat and 16% protein (Boyd, Arnbom, and Fedak 1993) compared with 47% fat and 17% protein during lactation. Thus, each year, females invest half as much energy in their skin as they do in their offspring. It seems fair to say that elephant seals look after their own skin.

Acquisition of Energy:
Movements and Behavior at Sea

Little is known about where southern elephant seals go after leaving breeding and molting beaches. But as this and other chapters in this volume demonstrate, we are starting to acquire information on foraging movements and diving behavior to complement the serial data on individuals gathered before and after they go to sea. Gathering such serial data is difficult at South Georgia because animals may breed and molt on widely separated islands and because of the large number of inaccessible breeding and molting sites available to animals that do return over several years.

Animals that bred and returned to the Husvik area to molt tended to arrive for the molt in the same order in which they left the beach after breed-

Fig. 20.4
The tracks of three postbreeding females obtained from Argos-compatible
transmitter packages. The cross-hatched area is the continental shelf. Locations
with a Location Quality Index (LQI) = 0 are shown as a cross and those with
LQI > 0 as a circle. All three seals swam southwest from South
Georgia to areas of the continental shelf.

ing. Their periods at sea ranged from 66 to 75 (mean = 72) days, during which time they gained 70 to 153 kg (mean = 107 kg, n = 8). Daily weight gains ranged from 1.1 to 2.3 (mean = 1.5) kg while at sea. Foraging success seems to vary significantly from individual to individual, even though the size range of animals recaptured was small (breeding beach departure weights of 340–457 kg). There was no relationship between size on leaving the beach and weight gain while at sea. The sample is, however, very small.

Four postbreeding female southern elephant seals at South Georgia were fitted with Argos compatible transmitter packages (Argos 1989) in November 1990 (McConnell, Chambers, and Fedak 1992). One failed when bitten by a copulating male. The movements of the remaining three animals are shown in figure 20.4. All traveled southwest to sites on the Antarctic continental shelf.

Seal 1 provided detailed information over 70 days. Its track was divided

Fig. 20.5
Typical dive depth profiles over 24 hours from seal 1 while  (A)  in phase 1 and
(B)  in phase 3. Four depth values were transmitted for each dive at intervals of
one-fifth of the duration of the dive. The maximum depth of each dive was also
transmitted and is shown as a dotted line in the figure. Preceding surface intervals
are represented by a horizontal line at zero depth. Breaks in the continuous plot
indicate missing data, due primarily to diurnal variation in satellite visibility. Note
the variability of dive depth and duration in phase 1 compared with phase 3.

into three phases based on location and movement. During phase 1 (23 days), it swam 1,845 km (average daily velocity [ADV] across the sea surface 0.93 m/sec.) to Livingston Island. In phase 2 (17 days), it hauled out at Livingston Island for 18 hours and then swam a further 805 km (ADV .55 m/sec.) to the southwest, following the continental shelf margin of the Antarctic Peninsula to a location 110 km west to Adelaide Island where water depth was 300 to 400 m. During this phase, it spent several periods of up to 12 hours at the surface. During phase 3 (29 days), it remained within 20 km of this location.

Seal 2 swam 1,420 km in 16 days (ADV 1.02 m/sec.) to Elephant Island where it hauled out. The transmitter failed one day later. It was captured, reweighed, and the transmitter removed at a molting site at South Georgia on February 8, 1991. Seal 3 swam 1,435 km in 16 days (ADV 1.03 m/sec.) to the continental shelf 110 km southeast of Elephant Island. It remained within a 60-km radius of this area for the next 4 days, after which the transmitter failed. It was sighted on January 19, 1991, at a molting site on King George Island.

Figure 20.5 provides a detailed view of dive profiles from seal 1 on two days typical of phases 1 and 3. Such dive-by-dive data were available from 65% of the animal's track: 6-hour dive summaries covered the entire period. We suggest that most of phase 1 was spent in transit to feeding

grounds, although the variability of maximum depths and dive durations suggests that there was some opportunistic feeding. Dives made during phase 3, when the animal remained in the small area off Adelaide Island, were shallower and less variable (fig. 20.6a) than during phases 1 and 2. The seal tended to swim directly to the bottom and remain there, sometimes swimming slowly, until it returned to the surface (fig. 20.6d). These dives involved less swimming activity than those in phases 1 and 2 (fig. 20.6e), yet they tended to be shorter (fig. 20.6b) with a smaller proportion of time spent underwater (fig. 20.6c). Between days 63 and 65, the seal moved 8 km north, 40 km west and then returned to its starting location. Over these 3 days, all dive parameters (fig. 20.6a–e) shifted to values similar to those in phase 1.

Seal 2 gained 141 kg over 78 days (equivalent to 1.8 kg/day) between breeding and molt. By analogy with female northern elephant seals whose weight gain is similar (Le Boeuf et al. 1989; Sakamoto et al. 1989), we estimate that South Georgia seals require 9 to 20 kg of prey per day; that is, 630 to 1,400 kg over the approximately 70-day interval between breeding and molt.

Our interpretation of the activity of seal 1 in phase 3 is of targeted benthic or demersal feeding. This is consistent with data from stomach samples taken on land in which cephalopods predominate. These cephalopods include species of demersal squid and benthic octopods (Rodhouse et al. 1992; Murphy 1914; Laws 1960; Clarke and MacLeod 1982). However, if, as we have demonstrated, feeding areas are far from breeding sites, these stomach samples may underrepresent the consumption of fish whose remains are retained in the stomach for shorter periods. Samples from seals hauled out on sites near foraging areas could help to asses this bias.

Dives in phase 3 were shorter by a factor of 1.5 than in phase 1, and thus more time was spent traveling to and from the surface than would be the case for longer dives to the same depth. The proportion of time spent underwater in these dives was 6% (81% vs. 87%) less on average than those in the first two phases. The following argument suggests that the seal was not merely resting during these dives. These animals spend 80 to 90% of their time underwater and might be better referred to as "surfacers" rather than "divers." If these were rest dives, there would seem to be little reason why they should be more frequently interrupted by traveling to and from the surface to breathe given that oxygen stores should last longer at rest.

But if these were feeding dives, why should they be shorter? Extended periods at the surface were rare during phase 3, implying that the processes of digestion and assimilation were combined with diving activity on a steady state basis and not delayed to breaks in diving activity. Assimilation of food is known to increase basal metabolic rate up to 1.7 times in another

Fig. 20.6
Dive parameters from seal 1 over the 70-day tracking period. Vertical lines demarcate
the three phases described in the text. Mean and standard error values, for each
phase, are shown for each parameter. The parameters refer either to individual dives,
to individual dives averaged over a day, or to 6-hour summary periods. Note that
during days 63 to 65 there was a temporary change in all dive parameters.
(A)  The maximum depth attained in each dive. Depth values are accurate to within
5 m at 100 m, to 50 m at 1,600 m. The maximum depth of approximately 910 m was
obtained during phase 1. The variability of maximum depths declined markedly
through phases 1 to 3 (SD = 171.2, 105.9, and 45.2). The constant upper limit of
depth in phase 3 corresponds to the depth of the seabed as
determined from Admiralty Charts.
(B)  The duration of each dive. Dive durations were shorter in phase 3 than
in phase 1 and less variable (SD = 2.5 and 5.4).
(C)  The percentage of time within each 6-hour summary period that the seal
spent underwater. This was least in phase 2, due primarily to extended periods on
the surface. The time spent underwater in phase 3, when we suggest the seal
was feeding, was less than during the transit in phase 1.
(D)  Index of dive depth profile squareness of each dive. Four depth readings
at intervals of one-fifth of the dive duration were transmitted in addition to the
maximum depth attained. The average of these four values are expressed as a
percentage of maximum depth. Dives were more flat bottomed in
phase 3 than in phase 1.
(E)  Daily averages of mid-dive swimming activity for each dive. A velocity turbine
provided an index of swimming activity over five equally spaced intervals within
each dive. The average of the middle three values is shown here to exclude descent
and ascent activity. Mid-dive activity was less in phase 3 than in phase 1.

phocid, the harp seal, Phoca groenlandica (Gallivan and Ronald 1981). We therefore suggest that the seal's aerobic dive limit was reduced during this phase, even though swimming activity was also reduced (fig. 20.6e), because the specific dynamic effect (SDE) (Kleiber 1961) of food assimilation increased the rate at which O₂ stores were used.

An additional component of this metabolic increase could also be the energy cost of warming the prey from sea temperature to body temperature, some 37°. The amount of heat this would require depends on the specific heat of the prey. If this is assumed to be equivalent to that of water (i.e., 4.19 J/g (and this assumption maximizes the likely effect since all prey will have specific heats less than this), then warming 20 kg of prey would account for the equivalent of about 10% of the standard metabolic rate of a 500-kg animal. However, this is not to say that warming food would require an increase in metabolic rate. This would depend on whether or not the animal was at or below the lower limit of its operational thermoneutral zone under the conditions prevailing while foraging. If it was not, it could decrease its heat loss via other avenues to the extent necessary to make up for the additional loss to the food. Given that SDE will cause a much larger increase in metabolism than 10%, the requirement for heating food in the stomach may not be important.

Interactions between Acquisition and Expenditure:
Linking Data from Land and Sea

Implications of Weight Change for Diving

Consideration of the huge weight losses of males (and, indeed, females as well) raises the question of how seals maintain effective control of buoyancy during diving throughout the annual cycle. Taking an extreme case, if a 3,000-kg male loses 1,000 kg while ashore and if most of the weight loss is lipid, then a thin animal would be much less buoyant than the same animal just prior to arrival on the breeding beach. Given that lipid is only 9 times the density of water (note that it is the density of the lipids used rather than blubber density that is important here), if the animal was neutrally buoyant on arrival, it could leave the beach 100 kg negatively buoyant, which amounts to a force equivalent to 3 to 5% of the weight of the animal.

Clearly, this is an extreme example, and the actual effect could be reduced, both because materials other than lipid are lost during the fast and because animals may come ashore positively buoyant and leave negatively buoyant, halving the imbalance. The imbalance could nonetheless amount to many kilograms upward or downward. While the imbalance in buoyancy would be much less in smaller animals, even small imbalances could be significant. Most human divers will be aware of the dramatic effect even small imbalances in buoyancy can have on maintaining their position in the water column and on the time they can remain submerged on a tank of air. This proportion of buoyancy change would amount to 4 kg extra of lead on a 70-kg human diver's weight belt, which would make for very hard work while operating free in midwater.

While an imbalance might have less energetic consequence during ascent and descent (losses in the energy required to swim up being partially repaid on the way down, or vice versa), it would have profound effects on the effort required to remain at a working depth in midwater or to remain on the bottom while buoyancy is positive. How might the animals compensate for the change in buoyancy? Changing the amount of gas taken down from the surface in the respiratory system would have little or no effect at the depths used by these species. While rocks and gravel are often found in the stomachs of elephant seals, it is hard to imagine enough being taken on board to compensate, and these might not be available in midocean. Could the physical properties (density and its relation to pressure and temperature) of fat stores of these seals be such that changes in the amount stored would not adversely affect buoyancy? Do animals change diving behavior, for example, switching from benthic to midwater foraging, as weight is regained to mitigate or take advantage of the effects of buoyancy changes? The questions remain open.

Foraging in Relation to Oceanography

The animals we tracked traveled more than 85 km each day for 16 to 23 days in a directed way to areas of the continental shelf. The use of distant foraging areas (associated with the Antarctic Polar Front, continental shelf margin, or ice edge) has also been inferred from water temperature data for elephant seals breeding on Macquarie Island (54°S, 157°E) (Hindell, Burton, and Slip 1991). Why do they adopt this strategy?

Elephant seal females are long-lived animals that invest large amounts of resources in a single pup each year over many years. They must, therefore, locate food reliably each year for many years in succession. Movement away from South Georgia may be explained by the fact that the local shelf area contains insufficient prey items to sustain the local elephant seal breeding population (McCann 1985). There may be an advantage, therefore, in adopting strategies that minimize the risk of yearly failure at the expense of the energetic costs of long-distance transits. The narrow Antarctic continental shelf, ice edges, and the Antarctic Polar Front are highly productive and attract many top predators (Ainley and De Master 1990) and fisheries (Itchii 1990). This contrasts with the open reaches of the Antarctic Ocean where concentrations of prey are both spatially and temporally variable and may be associated with unpredictable hydrographic conditions (ElSayed 1988). If an animal has the energy storage capability (as elephant seals, with their prodigious potential for blubber accumulation, do), the benefit of using distant foraging areas where food is reliably associated with readily relocatable oceanographic features, such as the continental shelf and the Antarctic Polar Front, may outweigh the costs of transport to these areas. That is, a long swim on an empty stomach may, in the long term, be more productive than pelagic meandering.

Variability in Size and Success

Yet, in spite of these capabilities, some animals gain more weight than others while at sea. And energy stores and the size at age of both males and females seem likely to be important to reproductive success. What factors of an animal's movements, experience, and foraging techniques are important in determining how successful it is?

Of particular importance is information on all aspects of the variability of foraging parameters: how they vary year to year in the same animal and change with age, sex, and size between animals. How do the patterns observed develop in naive animals when they first go to sea, and how does the variability of such early experience affect later development? No less important are the effects of the choice of breeding and molting sites in relation to near and distant oceanography.

Though the information to answer such questions is not available for

southern elephant seals, the techniques to do so are at hand. What is needed is their continued application and particularly their effective combination in land- and sea-based studies. These techniques will make new demands on the resolution, the accuracy, and, crucially, the availability of oceanographic data.

Acknowledgments

This work was carried out as part of the Sea Mammal Research Unit (Natural Environment Research Council) Open Oceans Programme in conjunction with the British Antarctic Survey and the University of Stockholm. We are particularly grateful to Neil Audley, Colin Hunter, Kevin Nicholas, Tim Barton, Ash Morton, and David Davies-Hughes for technical assistance and essential help in the field. We thank John Croxall for his usual unusually useful comments on the manuscript given good-naturedly with the usual short notice.

References

Ainley, D. G., and D. P. DeMaster. 1990. The upper trophic levels in polar marine ecosystems. In Polar Oceanography, Pt. B: Chemistry, Biology, and Geology , ed. W. O. Smith, 599–630. San Diego: Academic Press.

Argos. 1989. Guide to the Argos System . Toulouse: Argos CLS.

Boyd, I. L., T. Arnbom, and M. A. Fedak. 1993. Water flux, body composition, and metabolic rate during molt in female southern elephant seals (Mirounga leonina ). Physiological Zoology 66: 43–60.

Clarke, M. R., and N. MacLeod. 1982. Cephalopods in the diet of elephant seals at Signy Island, South Orkney Islands. British Antarctic Survey Bulletin 57: 27–31.

El-Sayed, S. Z. 1988. Seasonal and inter-annual variabilities in Antarctic phytoplankton with reference to krill distribution. In Antarctic Ocean and Resources Variability , ed. D. Sahrhage, 101–119. Berlin, Heidelberg, and New York: Springer Verlag.

Fedak, M. A., I. L. Boyd, T. Arnbom, and T. S. McCann. 1989. The energetics of lactation in southern elephant seals, Mirounga leonina , in relation to the mother's size. Abstract from the Eighth Biennial Conference on the Biology of Marine Mammals, Monterey, Calif., December 7–12, 1989.

Gallivan, G. J., and K. Ronald. 1981. Apparent specific dynamic action in the harp seal (Phoca groenlandica ). Comparative Biochemistry and Physiology . 69A: 579–581.

Hindell, M. A., H. R. Burton, and D. J. Slip. 1991. Foraging areas of southern elephant seals, Mirounga leonina , as inferred from water temperature data. Australian Journal of Marine and Freshwater Research 42: 115–128.

Huber, H. R. 1987. Natality and weaning success in relation to age of first reproduction in northern elephant seals. Canadian Journal of Zoology 65: 1311–1316.

Itchii, T. 1990. Distribution of Antarctic krill concentrations exploited by Japanese krill trawlers and minke whales. Proceedings of the NIPR Symposium on Polar Biology 3: 36–56.

Kleiber, M. 1961. The Fire of Life . New York: John Wiley and Sons.

Laws, R. M. 1956. The elephant seal (Mirounga leonina Linn). II. General, social, and reproduction behaviour. Falkland Islands Dependencies Survey, Scientific Reports 13: 1–88.

———. 1960. The southern elephant seal (Mirounga leonina Linn.) at South Georgia. Norsk Hvalfangst-Tidende 49: 466–476, 520–542.

———. 1984. Seals. In Antarctic Ecology , vol. 2, ed. R. M. Laws, 621–715. London: Academic Press.

Le Boeuf, B. J., Y. Naito, A. C. Huntley, and T. Asaga. 1989. Prolonged, continuous, deep diving by northern elephant seals. Canadian Journal of Zoology 67: 2514–2519.

Ling, J. K., and M. M. Bryden. 1981. Southern elephant seal, Mirounga leonina Linnaeus, 1758. In Handbook of Marine Mammals, 2, Seals , eds. S. H. Ridgway and R. J. Harrison, 297–327. London: Academic Press.

McCann, T. S. 1980. Population structure and social organisation of southern elephant seals, Mirounga leonina (L.). Biological Journal of the Linnaean Society of London 14: 133–150.

———. 1981. The social organization and behaviour of the southern elephant seal, Mirounga leonina (L.). Ph.D. dissertation, University of London, England.

———. 1985. Size, status, and demography of southern elephant seal (Mirounga leonina ) populations. In Sea Mammals in South Latitudes: Proceedings of a Symposium of the 52d ANZAAS Congress in Sydney—May 1982 , ed. J. K. Ling and M. M. Bryden, 1–17. Northfield: South Australian Museum.

McCann, T. S., M. A. Fedak, and J. Harwood. 1989. Parental investment in southern elephant seals, Mirounga leonina. Journal of Behavioural Ecology and Sociobiology 25: 81–87.

McCann, T. S., and P. Rothery. 1988. Population size and status of the southern elephant seal (Mirounga leonina ) at South Georgia, 1951–85. Polar Biology 8: 305–309.

McConnell, B. J., C. Chambers, and M. A. Fedak. 1992. Foraging ecology of southern elephant seals in relation to the oceanography of the Southern Ocean. Antarctic Science 4: 393–398.

Murphy, R. C. 1914. Notes on the sea elephant, Mirounga leonina (Linn.). Bulletin of the American Museum of Natural History 33: 63–78.

Reilly, J. J., and M. A. Fedak. 1990. Measurement of the body composition of living gray seals by hydrogen isotope dilution. Journal of Applied Physiology 69: 885–891.

Reiter, J., and B. J. Le Boeuf. 1991. Life history consequence of variation in age at primiparity in northern elephant seals. Journal of Behavioral Ecology and Sociobiology 28: 153–160.

Rodhouse, P., T. R. Arnbom, M. A. Fedak, J. Yeatman, and A. W. A. Murray. 1992. Cephalopod prey of the southern elephant seal, Mirounga leonina L. Canadian Journal of Zoology 70: 1007–1015.

Sakamoto, W., Y. Naito, A. C. Huntley, and B. J. Le Boeuf. 1989. Daily gross energy requirements of female northern elephant seal, Mirounga angustirostris , at sea. Nippon Suisan Gakkaishi 55: 2057–2063.

Worthy, G. A. J., P. A. Morris, D. P. Costa, and B. J. Le Boeuf. 1992. Molt energetics of the northern elephant seal. Journal of Zoology, London 227: 257–265.