Northern fur seal young - Canadian Science Publishing

4 downloads 0 Views 307KB Size Report
Abstract: We investigated size at birth, growth, and early survival of northern fur seals .... mass of northern fur seal pups on Robben Island during mid-.
843

Northern fur seal young: interrelationships among birth size, growth, and survival Alexander I. Boltnev, Anne E. York, and George A. Antonelis

Abstract: We investigated size at birth, growth, and early survival of northern fur seals (Callorhinus ursinus) from birth to weaning at Bering Island, Russia, over 8 breeding seasons from 1982 to 1989. One thousand and thirteen fur seals (565 males and 448 females) were measured in a longitudinal study and an additional 2697 animals were measured at birth. At birth, female pups were about 10% lighter and about 3% shorter than male pups. The coefficients of variation of mass (12.5 and 12.8%) and length (4.5 and 4.7%) were similar for the two sexes. We partitioned the lactation period into four time periods: 1, the perinatal period (ages 0–10 days); 2, the early development period (ages 11–40 days); 3, the period of intensive molting (ages 41–80 days); and 4, the preweaning period (ages 81–140 days). We investigated four measures of growth: absolute growth in mass (AGM) in grams per day, absolute growth in length (AGL) in millimetres per day, relative growth in mass (RGM) as a percentage per day, and relative growth in length (RGL) as a percentage per day. For both sexes, AGM was highest during period 4 (mean = 124.8 g/day, SE = 7.4 g/day, and mean = 109.6 g/day, SE = 6.8 g/day for males and females, respectively) and AGL was highest during period 2 (mean = 3.74 mm/day, SE = 0.18 mm/day, and mean = 3.42 mm/day, SE = 0.21 mm/day for males and females, respectively). RGM (mean = 1.06%, SE = 0.09%, and mean = 1.02%, SE = 0.11% for males and females, respectively) and RGL (mean = 0.53%, SE = 0.03%, and mean = 0.50%, SE = 0.03% for males and females, respectively) were highest during period 2 for both sexes. For both sexes, growth rates were slowest during the molting period. Sexual differences were detected in AGM in period 4 and for the combined data over periods 1–3. Sexual differences in AGL were detected for the combined data over periods 1, 3, and 4 only. No sexual differences in relative growth were found. Subsequent growth in mass and length was correlated with birth size. We found the greatest annual variation during the periods when growth was fastest. Condition indices were calculated using the allometric relationship between length and mass separately for neonates and pups older than 5 days. The condition indices at birth varied significantly annually. The condition index was lowest during the molting period (3). Animals that survived for at least 40 days were larger at birth and had a higher condition index than those that did not survive. In years of moderate or high pup survival rates, survival rates were higher in animals born later in the breeding season. Résumé : Nous avons étudié la taille à la naissance, la croissance et la survie au cours des premiers jours chez l’Otarie à fourrure (Callorhinus ursinus) de la naissance au sevrage, dans l’île de Bering, Russie, au cours de huit saisons de reproduction, soit de 1982 à 1989. La croissance en longueur a été suivie chez 565 mâles et 448 femelles et 2697 animaux additionnels ont été mesurés à la naissance. Les coefficients de variation de la masse (12,5 et 12,8%) et de la longueur (94,5 et 4,7%) sont semblables chez les mâles et les femelles. Nous avons morcelé l’allaitement en quatre périodes : 1, la période périnatale (de 0 à 10 jours); 2, le début du développement (de 11 à 40 jours); 3, période intense de mue (de 41 à 80 jours); 4, période de présevrage (de 81 à 140 jours). Nous avons examiné quatre mesures de la croissance : la croissance absolue de la masse en grammes/jour (AGM), la croissance absolue en longueur en millimètres/jour (AGL), la croissance relative de la masse en pourcentage/jour (RGM) et la croissance relative en longueur en pourcentage/jour (RGL). Chez les deux sexes, la croissance AGM est maximale au cours de la période 4 (moyenne = 124,8 g/jour, erreur type (SE) = 7,4 g/jour, et moyenne = 109,6 g/jour, SE = 6,8 g/jour pour les mâles et les femelles, respectivement) et la coissance AGL est maximale au cours de la période 2 (moyenne = 3,74, SE = 0,18 mm/jour, et moyenne = 3,42 mm/jour, SE = 0,21 mm/jour pour les mâles et les femelles, respectivement). La croissance RGM (moyenne = 1,06%, SE = 0,09%, et moyenne = 1,02%, SE = 0,11% pour les mâles et les femelles, respectivement) et la croissance RGL (moyenne = 0,53%, SE = 0,03%, et moyenne 0,50%, SE = 0,03% pour les mâles et les femelles, respectivement) sont maximales au cours de la période 2 chez les deux sexes. C’est au cours de la période de mue que les valeurs sont minimales chez les deux sexes. Il existe des différences sexuelles du taux de croissance AGM au cours de la période 4 et lorsque les périodes 1–3 sont combinées. Le taux AGL diffère chez les mâles et les femelles seulement lorsque les données des périodes 1, 3 et 4 sont combinées. Aucune différence due au sexe n’a été relevée dans les taux relatifs de croissance. La croissance en masse et la croissance en longueur après l’allaitement sont reliées à la taille à la naissance. La variation annuelle la plus grande se produit au cours des périodes où la croissance est le plus rapide. Les coefficients Received July 21, 1997. Accepted December 9, 1997. A.I. Boltnev. Kamchatka Research Institute of Fisheries and Oceanography, Naberedznaya 18, Petropovlosk, Kamchatka 683602, Russia. A.E. York1 and G.A. Antonelis.2 National Marine Mammal Laboratory, Alaska Fisheries Science Center, 7600 Sand Point Way NE, Seattle, WA 98115–0070, U.S.A. 1 2

Author to whom all correspondence should be addressed (e-mail: [email protected]). Present address: Southwest Fisheries Science Center, Honolulu Laboratory, Honolulu, HI 96822, U.S.A.

Can. J. Zool. 76: 843–854 (1998)

© 1998 NRC Canada

844

Can. J. Zool. Vol. 76, 1998 d’embonpoint, établis à partir de la relation allométrique entre la longueur et la masse, ont été calculés séparément pour les nouveau-nés et les petits de plus de 5 jours. Le coefficient d’embonpoint à la naissance subit des variations annuelles significatives. C’est au cours de la mue (période 3) que le coefficient d’embonpoint est le plus faible. Les otaries qui survivent plus de 40 jours sont des animaux plus gros à la naissance et ils ont un coefficient d’embonpoint plus élevé que les animaux qui ne réussissent pas à survivre. Au cours des années où la survie des jeunes est moyenne ou élevée, ce sont les animaux nés plus tard au cours de la saison de reproduction qui ont les taux de survie les plus élevés. [Traduit par la Rédaction]

Introduction Growth is an important factor in the life history of animals and many vital parameters are related to it, such as age at maturity, survival rate, and reproductive effort (SchmidtNielsen 1984; Harvey et al. 1990; Promislow and Harvey 1990; Roff 1992). Mammalian growth has been the subject of numerous studies and several monographs summarize these investigations (e.g., Cross 1972; Brody 1945; Mina and Klevezal’ 1976). The importance of food resources is recognized as one of the key factors influencing individual and population growth (Wynne-Edwards 1962; Andrewartha and Birch 1984). Assuming that growth processes integrate over an animal’s physical and biological environment, neonatal growth could be an important indicator of population stress (e.g., Eberhardt and Siniff 1977; Zach 1988; Gerrodette and DeMaster 1990; Fowler and Siniff 1992). In an effort to identify factors possibly influencing population growth of seabirds and seals, scientists working in the Antarctic are attempting to use such indices as growth of offspring for indirect evaluation of prey availability (Croxall et al. 1988). Several studies on the development of young pinnipeds have examined the relationship between food availability and the ability of lactating females to optimize the growth of their offspring (Reiter et al. 1978; Mattlin 1981; Doidge et al. 1984; Kovacs and Lavigne 1986; Trillmich 1986; Costa et al. 1989; Lunn et al. 1993; Bowen et al. 1992; Gentry1998). This relationship was clearly demonstrated during the 1982–1983 El Niño event, when changes in prey resources were associated with declines in mass, growth, or survival of neonates of several species of pinnipeds (Trillmich and Ono 1991). The theory of sexual selection (Trivers and Willard 1973; Maynard-Smith 1980; Clutton-Brock 1991) predicts high maternal investment in males of polygynous species. Thus, all otariids are ideal species in which to investigate this idea (Trillmich 1986). Polygynous pinnipeds produce sexually dimorphic young at birth (Scheffer 1981; Kolesnik and Timofeyeva 1980; Mattlin 1981; Little et al. 1987; Boyd and McCann 1989; Le Boeuf et al. 1989; McCann et al. 1989; Merrick et al. 1995). Differences in the rate of growth in mass of otariid young during lactation have been cited as evidence of differential maternal investment (Doidge et al. 1984; Kerley 1985; Trillmich 1986; Lunn et al. 1993; Goldsworthy 1995). We have found no reports of differential growth in length among otariid young. The level of maternal investment in offspring should influence their growth and survival. Evidence of a direct effect of size on survival was demonstrated by Baker and Fowler (1992), who showed that the size of male northern fur seal (Callorhinus ursinus) pups in late August (when they are 1–2 months old) is positively correlated with their survival to 2 years of age.

The timing of events in the life history of the northern fur seal is important for understanding the results of our analyses. This chronology has been summarized by several authors (e.g., Bartholomew and Hoel 1953; Peterson 1968; Marakow 1974; York 1987; Gentry 1998). Northern fur seals spend most of the year at sea over a large area in the North Pacific Ocean and adjacent seas. During summer they go ashore to breed in colonies in Russia, Alaska, and California. From mid-May through early June, adult males arrive on the breeding grounds, where they establish territories that they maintain through early August. Pregnant females arrive from midJune through early August. They give birth within 1 or 2 days to a single pup and mate within the following week. In Alaska and Russia the birthing peak occurs in early July and the breeding peak in mid-July. The postpartum female then begins a series of cycles: feeding at sea for 5–12 days followed by nursing on land for 2–3 days. Pups are waxy and black when they are born and most have molted to a silver adult-like pelage by September. Fur seal pups begin molting very shortly after birth; the period of intensive molt lasts from ages 1.5 to 2.5 months, with a peak at about 2 months (Scheffer 1961; Scheffer and Johnson 1963; Bauer et al. 1964; Surkina 19853). Pups are weaned during late October to November (Gentry and Holt 1986), when they are 4–5 months old. Kolesnik and Timofeyeva (1980) studied the growth in mass of northern fur seal pups on Robben Island during midJuly through late September from cross-sectional data. Calambokidis and Gentry (1985) and Gentry (1998) studied the growth in mass of known-age fur seal pups from birth to 50 days on the Pribilof Islands. The goals of this paper were to (i) study growth in mass and length of known-age pups during the lactation period, (ii) evaluate the influence of molting, size at birth, and birth date on pup growth, and (iii) study sexual differences in growth rates. We analyze neonatal mass and length information collected on Bering Island, Russia, to produce a growth curve for northern fur seal pups. Interannual variations in the size of pups at birth and their growth are compared for 8 reproductive seasons. We examine the allometric relationship between mass and length of pups and use this information to describe inter- and intra-annual changes in their condition. We also investigate which factors appear to influence growth and survival before weaning.

Methods Mass and length measurements were collected from northern fur seal 3

R.M. Surkina 1985. Morphological and historical investigation of skin and pelage of male northern fur seals on the Commander Islands during the lactation period. [In Russian.] Unpublished report, Academy of Science, Zoological Institute, Ukraine, USSR. © 1998 NRC Canada

Boltnev et al. pups on North Rookery, Bering Island (52°N, 161°W), Russia, from 1982 to 1989; all measurements were collected between June 20 and November 19, although the range of collection dates varied from year to year. Birth measurements and sex information were collected during the June and July pupping seasons, using roving two-person observation blinds (approximately 1 × 1.5 × 2 m, 40–50 kg). In contrast to all previous studies on the growth of fur seals, these mobile blinds provided access to almost all pups on the rookery and gave researchers protection from attacks by territorial males while pups were measured. Pups were captured by hand, noose pole, or net and brought into the blind, where they were measured and tagged and their sex was determined. Pups were considered newborn (less than 1 day old) if they had a fresh umbilicus and their flippers had not yet darkened to black, or if dark green fetal excrement was present (Boltnev 1994). They were then placed, without the placenta, in a plastic bucket or canvas bag and weighed, wet or dry, with a spring scale to the nearest 0.2 kg. Dorsal standard length was measured (McLaren 1993) to the nearest 1.0 cm by holding the pup belly down, stretched on a flat board that was marked in 1-cm increments. The pup’s nose was held against a block of wood attached to one end of the measuring board while the body was straightened by pulling gently on the rear flippers. Prior to the pups’ release at first capture, each foreflipper was tagged with a numbered Monel steel tag (Antonelis 1992), and pups were returned to their mother within 2–3 min after capture. Subsequent measurements of pups were collected from the roving blind or during periodic measurements of groups of pups, when marked and unmarked pups were recaptured and individually measured and released. All dead tagged pups (281 out of 3690 tagged pups with known birth date and birth size) encountered during field or census work were collected; the known date of last measurement of these pups was a lower bound on their life-span. Analysis of data To standardize the interannual comparison of sizes at birth, we analyzed measurements collected between 5 and 9 July, near the peak of pup production (Marakow 1974). Birth masses and lengths were compared for differences between years and sexes using a two-way analysis of variance; multiple comparisons were made with a Tukey q (HSD) test (Zar 1974). We used the “successive differences” method described by Schwertman and Heilbrun (1986) to compare the longitudinal growth of pups from birth to the time of weaning. This method uses differences in consecutive pairs of observations for all subjects having two or more measurements and is especially applicable to growth data for which sampling was not uniform (either the number of sampling occasions or the date was the same) for each subject. The main assumption is that for each subject, growth is a piecewise linear function; that is, within a certain time frame, growth is linear but there may be different slopes in different time frames and the slopes may be different for different subjects. Parameters for all models were estimated as general linear models using the statistical software SPlus (Becker et al. 1988). Assessment of statistically significant differences in growth rate between different groups (e.g., between males and females), differences in rates over time, or the effect of size at birth on subsequent growth were done using likelihood-ratio tests with degrees of freedom adjusted using the Geisser and Greenwood method (Geisser and Greenwood 1958; Schwertman and Heilbrun 1986). This adjustment in degrees of freedom was carried out because not all measurements were independent and it was important to adjust for the effects caused by correlations in the size of individuals from one sampling occasion to another. Absolute (or daily) growth in mass (AGM; in grams per day) and length (AGL; in millimetres per day) were determined from the differences in mass and length measurements obtained on successive capture dates. Relative growth in mass (RGM) and relative growth in length (RGL) as a percentage per day were determined from the differences in logarithms of mass and length. We estimated growth rates

845 during four age periods, 1 (0–10 days), 2 (11–40 days), 3 (41–80 days), and 4 (81–140 days), corresponding to the perinatal, early development, molting, and preweaning periods. In our models, we assumed that growth was a linear function within each of these periods. These periods were defined using the results of previous studies on molting, which placed the period of intensive molt for northern fur seals between ages 40 and 80 days, and from data collected during the first field season which suggested that during the period from birth to the end of the mother’s first feeding trip, approximately 10 days, pups maintained their size and did not appear to be growing noticeably. We also investigated the sensitivity of our results to the definitions of the periods by varying their endpoints. To estimate growth rates, the response variable, e.g., absolute or relative change in mass or length, was regressed on the length of time in each of the four periods spanned by the particular change. The effect of mass or length at birth on subsequent growth was determined by comparing the sex-specific differences in growth rates of “small” (smallest third), “average” (middle third), and “large” (largest third) newborn animals using likelihood-ratio tests. When many statistical tests are performed at a particular level (e.g., Fisher and van Belle 1993), all the tests are not simultaneously valid at that level. We used the Bonferroni method, a conservative approach, to determine the level at which a number of statistical tests was significant (Fisher and van Belle 1993). Parameters were estimated by regressing the mass or length difference (or the differences in their logarithms) on the number of days in each period spanned by the observation. For example, if an animal was weighed at ages 0 (5 kg), 19 (6 kg), and 120 (16 kg) days, then the first mass gain is 1 kg, which spans 11 days in period 1, 8 days in period 2, and 0 days in periods 3 and 4. The second mass gain of 10 kg spans 21 days in period 2, 40 days in period 3, and 40 days in period 4. In this example the pup was not measured during period 3, but information is available about its growth during that period because we observed the animal’s size over a time period that includes period 3. The parameters describing the allometric relationship between mass and length for known-age pups (M = cLk ) were estimated by regressing the natural logarithms of mass and length on each other. From this relationship we could predict the expected or average mass at an observed length.We used this relationship to define a condition index, calculated by dividing the observed mass by the predicted mass in the allometric equation. To avoid biases caused by unequal samples of individual pups, and since our sample sizes were ample, we used only one randomly selected sample (mass and length) from each pup to estimate the parameters of the allometric relationship. We modeled survival to age 40 days using a general linear model with survival as a binomial response (McCullagh and Nelder 1989) and investigated the effects of size at birth, condition, birth date, and birth year on survival.

Results The number of individual fur seals available for our longitudinal study over 8 field seasons was 1013 (565 males and 448 females). An additional 2697 animals were measured at birth. Summaries of sample sizes, means, and standard errors of all measurements of mass and length by 10-day period are given in Table 1. Size at birth Animals were measured during the peak of breeding during 6 field seasons (Fig. 1). The mean mass of male pups (6.13 kg, SE = 0.044 kg) was about 10% greater (P < 0.0001) than that of female pups (5.55 kg, SE = 0.033 kg); yearly differences were consistent for both sexes. The range of annual means of birth mass of females (5.26–5.90 kg) was somewhat narrower than that of males (5.67–6.43 kg); however, the ranges over© 1998 NRC Canada

846

Can. J. Zool. Vol. 76, 1998

Table 1. Mean mass and length with standard deviations and sample sizes. Mass (kg) —————————————————————————————— Females Males ————————–—— ————————— Age (days) Mean SD n Mean SD n Birth 1–10 11–20 21–30 31–40 41–50 51–60 61–70 71–80 81–90 91–100 101–110 111–120 121–130 131–140

5.51 5.88 6.54 6.98 7.41 8.50 9.40 9.80 10.56 11.08 12.12 12.98 14.29 14.58 16.77

0.74 0.69 0.93 0.94 0.80 — — — 1.53 1.84 1.50 2.19 2.39 2.21 1.90

1349 155 81 79 37 1 1 1 9 34 45 96 96 42 6

6.08 6.45 7.22 8.00 8.65 9.30 11.00 10.60 12.04 12.86 13.84 14.97 15.88 17.21 20.08

0.81 0.92 1.24 1.27 1.30 1.44 0.50 — 0.97 3.17 2.04 2.19 2.83 2.78 4.27

1347 215 91 89 59 13 3 1 7 29 45 110 118 69 22

Length (cm) ———————————————————————— Females Males —————————— –————————–—— Mean SD n Mean SD n 62.47 63.39 66.65 68.49 71.05 75.00 81.00 74.00 82.67 80.62 81.13 82.07 82.93 84.71 84.83

2.87 3.22 2.74 3.02 4.27 — — — 5.36 2.94 3.47 3.24 5.03 3.42 2.14

1348 154 81 77 37 1 1 1 9 34 45 95 95 42 6

64.60 65.24 68.74 71.59 64.33 77.08 85.00 79.00 84.43 83.03 85.07 85.68 87.35 89.03 93.00

3.04 3.14 3.46 3.57 4.20 4.43 2.00 — 3.36 6.85 4.51 3.60 4.68 3.87 3.80

1345 215 90 88 58 13 3 1 7 29 45 110 118 69 20

Note: Animals were measured at the north rookery, Bering Island, Commander Islands, Russia, during summer field seasons in 1982–1989.

Fig. 1. Mean mass (kg) at birth of northern fur seal pups born during the peak of births (5–9 July) at North Rookery, Bering Island, Russia, in 1982, 1983, 1985–1987, and 1989. Estimated 95% confidence intervals are shown.

lapped, and the largest annual birth mass for females (in 1983) exceeded the smallest annual birth mass for males (in 1989). The birth mass of both sexes declined during 1983–1989 (P < 0.05). In 1983, the mean mass of males and females was significantly (P < 0.05) greater than in all other years, and mass was lower in 1989 (P < 0.05) than all other years. The variance of birth mass was about 20% higher in males than in females, but the coefficients of variation (CVs) were similar: 12.5 and 12.8% for males and females, respectively. For animals born during the peak of breeding, the mean length of male pups (64.8 cm, SE = 0.38 cm) was about 3% greater (P < 0.001) than that of female pups (62.7 cm, SE = 0.44 cm). The range of yearly mean birth lengths for females

(60.9–63.4 cm) was narrower than for males (62.3–66.5 cm). Yearly differences in birth length were not always the same for males and females. Length at birth increased between 1982 and 1986 and then declined from 1986 to 1989 (Fig. 2). The mean length of females was significantly less (P < 0.05) in 1982 and 1989 than in all other years, and the mean length of males was significantly less (P < 0.05) in 1989 than in all other years. There was no difference (P > 0.05) in the mean length of females at birth from 1983 to 1987. The variance of length was about 15% higher in males (P = 0.097), but CVs were similar (4.5% for males and 4.7% for females). The CVs of length were significantly less than those of mass (P < 0.001). © 1998 NRC Canada

Boltnev et al.

847

Fig. 2. Mean length at birth (cm) of northern fur seal pups born during the peak of births (5–9 July) at North Rookery, Bering Island, Russia, 1982, 1983, 1985–1987, and 1989. Estimated 95% confidence intervals are shown.

Fig. 3. Mass at age with lowess (Cleveland 1979) fit superimposed. Males are heavier than females during the whole lactation period.

Size at age A comparison of mass at age (Fig. 3) and length at age (Fig. 4) for males and females shows that males were consistently larger than females and that daily mass and length gains were similar for both sexes. A spurt in length gain took place between ages 10 and 40 days (period 2), when growth in mass appeared to be slower and more constant (compare Figs. 3 and 4). Below we give estimates of growth during each of the four periods. We studied the sensitivity of the ranges for the four periods in estimating growth rates by changing the definition of the molting period, extending it from 30 to 90 days of age. This change did not significantly influence the estimated growth rates. Absolute growth in mass: daily gain in grams per day The daily mass gain (AGM; Table 2) followed a similar pattern over time for males and females. Growth was slow but

highly variable in the perinatal period, increased during period 2, decreased somewhat during period 3 (the period of intensive molting), and increased sharply before weaning (period 4). For both sexes, AGM was significantly higher during period 4 (P < 0.01) than in periods 1–3 and during period 2 (P < 0.04) than in period 3. The estimated daily mass gain for males always exceeded that for females, but during each of the first three periods, none of these differences were significant (P > 0.05). The daily mass gain averaged over periods 1–3, however, was greater (P < 0.001) for males (70.8 g/day, SE = 2.5 g/day) than for females (57.5 g/day, SE = 2.3 g/day). In period 4, males gained more mass than females (P = 0.039). The standard errors of the estimates of daily mass gain decreased over periods 1–4 for both sexes. There were no annual differences in AGM for period 1, 2, or 3, nor in the daily mass increase averaged over periods 1–3 (P > 0.096 in all cases). In period 4 we had sufficient data to © 1998 NRC Canada

848

Can. J. Zool. Vol. 76, 1998

Fig. 4. Length at age with lowess fit superimposed. Males are longer than females during the whole lactation period.

Table 2. Growth in mass (AGL, g/day) and length (AGL, mm/day) of northern fur seal pups at North Rookery, Bering Island, Commander Islands, Russia, in 1982–1989.

Period 1 2 3 4

(0–10) (11–40) (41–80) (81–140)

Males ——————————————— Mass Length ——————— —————— Mean SE Mean SE

Females ————————————–——— Mass Length ———–——— ———–——— Mean SE Mean SE

55.19 83.69 62.83 124.82

35.87 70.84 51.90 109.64

19.78 11.40 8.68 7.36

1.27 3.74 1.19 1.60

0.32 0.18 0.14 0.12

18.71 11.45 7.55 6.77

1.28 3.43 1.08 1.41

0.35 0.21 0.14 0.13

Note: Numbers in parentheses show age in days.

compare the daily mass increases in 1987 and 1989. Males grew faster in 1989 (P = 0.02); there was no difference for females (P = 0.077). Absolute growth in length: daily gain in millimetres per day Daily growth in length (AGL) followed similar patterns for males and females (Table 2). Both sexes grew about 3 times faster during period 2 than in all other periods (P < 0.001). The daily length gain in periods 1, 3, and 4 was significantly smaller than in period 2. There was no significant difference in AGL between males and females (P > 0.05) within all four growth periods, although AGL was never less in males than in females. The daily length gain averaged over periods 1, 3, and 4 was greater (P < 0.01) for males (1.42 mm/day, SE = 0.05 mm/day) than for females (1.26 mm/day, SE = 0.06 mm/day). During period 2 we found annual differences in AGL (P < 0.05), which ranged from about 2.5 to about 4.5 mm/day for males and from 2.0 to 4.0 mm/day for females. The yearly differences were not consistent between males and females. The smallest AGL values were observed in 1987 and 1989

(both sexes) and the highest in 1982 (both sexes), 1984 (females only), and 1983 (males only), with intermediate values in 1986 for both sexes, in 1983 for females, and in 1984 for males. For all years except 1983, the daily length gain in period 2 did not differ between males and females (P = 0.3); in 1983, however, it was significantly greater (44% higher) for males (P < 0.01) than for females. The standard errors of the estimates of daily length gain decreased over periods 1–4 for both sexes. Relative growth in mass The patterns of RGM, measured as daily percent increase in mass, were similar for males and females. Values were highest in period 2 and lowest in period 3, with intermediate values in periods 1 and 4. RGM was significantly greater in period 2 than in periods 3 and 4 for both sexes (P < 0.05 in all cases). For females, RGM was significantly greater in period 4 than in period 3 (P < 0.001). There were no significant differences (P > 0.05) in RGM between periods 1 and 4 or between periods 1 and 3. In period 4, RGM was higher in 1987 than in 1989 for both sexes (P < 0.01 in all cases). In period 2, RGM was lower in 1987 than in all other years (P < 0.05). © 1998 NRC Canada

Boltnev et al.

849 Table 3. Relative growth in mass and in length (percent per day) of northern fur seal pups at North Rookery, Bering Island, Commander Islands, 1982–1989.

Period 1 2 3 4

(0–10) (11–40) (41–80) (81–140)

Males ——————————————— Mass Length ——————— —————— Mean SE Mean SE

Females ————————————–——— Mass Length ———–——— ———–——— Mean SE Mean SE

0.95 1.06 0.64 0.80

0.81 1.02 0.52 0.88

0.16 0.09 0.07 0.06

0.21 0.53 0.15 0.18

0.04 0.03 0.02 0.02

0.17 0.11 0.07 0.06

0.22 0.50 0.13 0.19

0.05 0.03 0.02 0.02

Note: Numbers in parentheses show age in days.

Table 4. The significance of the effect of birth size on four measures of growth (AGM, mass in grams per day; RGM, percent mass per day; AGL, length in millimetres per day; RGL, percent length per day).

Period

Birth mass ———————–———————— AGM RGM AGL RGL

1 2 3 4

0.311 0.201 0.039 0.050