Birth timing after the long feeding migration in elephant seals

Northern elephant seals migrate long distances from feeding grounds to raise pups during a brief period on breeding beaches. Since gestation sets a parturition date months in advance, timing of the arrival must be precise. We used satellite-tracked animals to examine this timing, establishing arrival and birth dates in 106 migrating females and estimating how far they traveled in the days just before birth. Females arrived a mean of 5.5 days prior to birth (range 1-11, sd=1.6), and females arriving later in the breeding season cut that pre-birth interval by 1.8 days relative to early arrivers. There was no correlation between female body condition, nor female age, and the pre-birth interval. The last 15 days prior to birth, animals traveled as far as 1465 km. Those furthest from the colony traveled > 100 km per day, three times faster than animals near the colony at the same time. Despite migrations covering several thousand kilometers while pregnant, female elephant seals were able to time their arrival within 6 days, swimming steadily at high speed if needed. This allows them to maintain a precise annual cycle for many years consecutively.

Reproductive phenology, or the timing of reproduction, is a core feature of life history in 2 animals and plants. A prominent phase of the reproductive schedule in migratory 3 animals is the transition between migration and reproduction. In birds, the phrase 4 'arrival biology' is used to describe this transition, known variously as the pre-chick, 5 post-arrival, pre-breeding, or pre-laying period [ [1][2][3][4][5]]. The transition is a small part of 6 the annual cycle, but details of its timing deserve attention. From a distant location, 7 animals must initiate a long migration so that they arrive at the breeding ground on a 8 precise schedule [ [6,7]]. 9 Mammals are also migratory, and marine mammals travel long distances between 10 feeding and breeding. Elephant seals (Mirounga spp.), for example, spend most of their 11 lives hunting for fish and squid in remote oceans but then migrate thousands of 12 kilometers in order to birth and raise pups on land [ [8][9][10][11]]. Unlike birds, gestation 13 begins months in advance, setting a parturition date and requiring precise timing of 14 arrival at the breeding colony: too early means wasted foraging time, while too late 15 would be fatal to the newborn. Yet females maintain this precision year after year and 16 give birth on a consistent annual cycle [ [12]]. Our broad goal is understanding what 17 limits female reproductive success and how individuals maximize foraging time and 18 pupping success [ [13][14][15][16][17][18]], including the birth timing and whether it limits fecundity. 19 Here we take advantage of a sample of female northern elephant seals (M. 20 angustirostris) that were tracked by satellite during their migration prior to parturition. 21 Satellite-derived locations document movements at the end of the migration as well as 22 the exact arrival time at the colony, while direct observations of pups establish 23 parturition dates. With those observations, we can estimate the time interval between 24 arrival and birth and ask whether 1) more experienced mothers, 2) mothers in poor 25 condition, or 3) late arriving mothers managed to shorten their pre-birth intervals. We 26 also calculated the distance traveled in the last two weeks of the migration to examine 27 females' ability to control arrival time, asking whether animals further from the colony 28 traveled at a higher speed. 29

30
Breeding and the annual cycle 31 Northern elephant seals (Mirounga angustirostris) breed on remote beaches from Baja 32 California to Vancouver Island. The largest colonies are in central California between 33 32 • to 38 • N. latitude [ [19,20]]. Females spend 31 days on the colonies in winter where 34 they birth and nurse single pups, then copulate and wean the pup to depart on long 35 foraging migrations (Fig 1). The females are far from shore for 2-3 months, then return 36 to the colony in April or May to molt. At about this time, the fertilized ovum implants 37 in the uterine wall, so gestation takes place during the following seven-month foraging 38 migration in the summer and autumn [ [21,22]]. Recorders were deployed on individuals to document movements while at sea [ [8,9]].

47
Instruments were attached prior to foraging migrations and removed immediately 48 thereafter [ [24]]. While the animals were sedated, detailed morphometric measurements 49 were collected in order to estimate body composition as percent fat [ [8,13,25,26]].  Periods at sea are in blue, on land in brown. The pre-birth interval is marked with an asterisk. these cases, the time of birth was observed with a precision of 24 hours, because 62 observations of pups were confined to daylight hours during the short winter days.  The distance from the colony was calculated 15, 10, and 5 days prior to parturition 80 from the ARGOS satellite locations, providing a female was still at sea. For all 106 81 females, we calculated the net travel toward the colony between day 15 and day 10, 82 which is the colony-distance on day 15 minus the colony-distance on day 10; it was 83 divided by 5 to express net daily travel.

85
Arrival and the pre-birth interval 86 Females arrived between 28 Dec and 8 Feb, with a mean of 15 Jan (SD=8.3 d). They 87 gave birth D = 1-11 days later, between 3 Jan and 9 Feb (mean 21 Jan, SD=7.9 d).

88
The 1-day and 11-day delays were outliers, however, since every other interval D was 89 3-8 days (Fig 2). The sample mean of D was 5.49 d, with a standard deviation of 1.53 d. 90 The Gaussian model fit the distribution of D better than gamma and Laplace 91 distributions, though even the Gaussian failed to account for narrow peak at 4-6 days 92 (Fig 2). The Laplace distribution captured the steep peak but had a poorer fit in the 93 tails. Variation in the pre-birth interval 95 There was a significant negative relationship between arrival date and the pre-birth 96 interval (Fig 3). The mean interval for females arriving on 1 Jan was 6.3 days, 97 decreasing to 4.5 days for females arriving on 1 Feb. There was no relation between 98 female age and the pre-birth period (r 2 = 0.02, p = 0.18), nor between a female's 99 condition (body fat) and the interval (r 2 < 0.01, p = 0.64).

100
The shortest pre-birth interval was a single female with a pup one day after arriving 101 (Figs 2, 3). She arrived just after midnight and was seen without a pup on the first day, 102 then with a pup the following day, so the birth happened 15-30 hours after arrival. She 103 was also the latest of all the females to arrive, on 8 Feb, 24 days later than the average. 104 All other females had at least a 3-day delay between arrival and parturition. records. Arrival is given in days since 1 Dec, so day 32 is 1 Jan, day 63 is 1 Feb. The regression coefficient is −.059 (p < 0.001, r 2 = 0.10). The solid black line is the regression; the light gray curve connects the mean delay in 5-day intervals (< 35, 35-39, 40-44 ... ≥ 55). Points were moved slightly at random to reveal where multiple points coincided.
The single female who waited 11 days, 3 days longer than any other (Fig 2), was 106 observed 16 times over those initial 11 days, by many different observers, never with a 107 pup. She was then seen only four times with a pup, and the pup was marked during the 108 procedure to retrieve her satellite tag. Subsequently, for her final 20 days on the colony, 109 she was separated from her marked pup. She arrived on 1 Jan, well before average but 110 not the earliest (Fig 3).

111
Migration rate prior to birth 112 Fifteen days prior to parturition, when all females were still at sea, the mean distance 113 from the colony was 681 km; the distribution of distances was nearly symmetric, and 114 the median (650 km) was only slightly lower than the mean. There were 11 females 115 > 1000 km away, and the furthest was 1465 km. At 10 days preparturition, the mean 116 distance was 369 km, with the furthest female 941 km away. At five days prior, nearly 117 half the females were on the colony; the rest were an average of 130 km away, and the 118 furthest 448 km. Between 15 and 10 days before parturition, females traveled an 119 average of 316 km (63 km day −1 ), and there was a strong positive correlation between 120 distance at day 15 and how far they traveled (Fig. 4).

122
Female elephant seals arrived on the colony 5.5 days before giving birth, ranging from 123 1-11 days. Given how far they traveled over the seven months beforehand and the Figure 4. Pre-birth distance traveled. Distance from the colony 15 days before birth (x-axis) versus net travel per day toward colony in the next five days (y-axis). The slope is +0.067 km faster daily travel for every km further from the colony (r 2 = 0.68, p < .001, N = 106).
Several of the females traveled > 80 km per day in the last two weeks. The observed 126 correlation between speed of travel and distance from the colony suggests the females 127 knew where they were and how soon the pup was due. The fact that most females 128 timed their birth 3-8 days after such a lengthy trip further demonstrates the precision of 129 their cycle.

130
Selection should favor females who can reduce the delay between arrival at the 131 breeding site and birth, because the delay wastes both foraging time and nursing time. 132 The same selection is expected in other long-distance migrants. In cases where the delay 133 has been quantified, caribou gave birth 4 days after arrival [ [28]] and humpback whales 134 17 days [ [29]], while pied flycatchers laid eggs 10 days after arrival, black-throated blue 135 warblers 20 days [ [6]], and salmon 20 days [ [30]]. White storks, though, waited a much 136 longer 59 days [ [31]]. There are important contrasts among these species, however, since 137 gestation occurs throughout the migration in the large mammals, but birds develop eggs 138 after arrival. Shorebirds in Greenland offer a well-studied example of the latter, because 139 the eggs and young are provisioned entirely by local feeding [ [3,32,33]], whereas in 140 elephant seals and whales, all provisioning -gestation through nursing -is supported by 141 foraging at distant feeding grounds. The birds are thus under selection to adjust the 142 pre-breeding delay in response to food availability on the breeding ground [ [4,34,35]], a 143 factor irrelevant in elephant seals and whales. Overall, the short delay in elephant seals 144 and caribou, 4-5 days, can be traced to gestation while migrating. Birds delay laying  gestation to nursing inside one year because failing to do so would mean missing a year 160 of reproduction in order to restore the seasonal rhythm [ [38]]. Maintaining the annual 161 cycle is thus crucial to a female's long-term reproductive output [ [12]]. Our next goals 162 include elucidating the timing of the full cycle, including reproduction as well as the 163 annual molt (Fig. 1), and how lifetime phenology of individual females affects lifetime 164 reproductive success.