Reproductive ecology of female South American fur seals at Punta San Juan, Peru. By Maria Patricia J. Majluf Chiok King's College UNIVERSITY LIBRARY CAMBRIDGE This dissertation is submitted to the University of Cambridge in application for the degree of Doctor of Philosophy August 1987 Frontispice South American fur seal female and her pup " ... Extraordinary uniformity is , after all , the outstanding oceanic feature of the Peruvian littoral. The very fact that relatively slight fluctuations toward the border line of a distinct region so profoundly affect the whole life association of the Humboldt Current , is, in its-elf, evidence that this life must have become thus inflexibly adapted through the long duration of sustantially invariable conditions. In the cosmic scheme the annihilation of a myriad sea birds and lesser organisms is a trifling incident ; almost before the effect can be appreciated, the circulation changes, the cold, upwelling waters are reinvaded by new hordes; and within a few days or weeks the old order reigns again, heedless of the Damoclean sword, from Mollendo to Point Parina ... " R.C. Murphy 1925 Bird Islands of Peru Preface I declare that this dissertation is the result of my own research. Where work has been collaborative due acknowledgement has been given . This dissertation does not exceed 80,000 words and has not been submitted in whole or in part to any other University for consideration of a higher degree. ABSTRACT The breeding patterns and foraging ecology of female South American fur seals (Arctocephalus australis) were studied at Punta San Juan, Peru from January 1983 to March1987. The particular interest of fur seals in Peru is that they face high air temperatures while breeding on land and unpredictable fluctuations in their food supply due to El Nit'lo events. This is in contrast to most other fur seals which breed in temperate to sub-polar environments and have a more predictable food supply . Scat analysis showed that Peruvian fur seals feed mainly on small epi- and meso-pelagic fish, mainly anchovetas (Engraulis ringens) . Dive records showed that they hunt at night when their prey are closer to the surface. During El Nino years, when anchovetas occur at greater depths than usual, fur seals make a greater number of deep dives, thus having to spend more energy to obtain prey than in non-El Nit'lo years . Low prey availability also result in mothers having to spend longer periods than usual foraging at sea and young pups dying of starvation during their mothers' extended abscences. The breeding pattern of fur seals in Peru differs little from that of other species of fur seal. Breeding occurs within a defined season from mid-October to late December. Males are larger than females and defend territories on land to prevent intruders from having access to females . Females produce a single pup. Mating takes place 8-10 days after pupping . From then on, females divide their time between suckling their young on land and foraging at sea. In contrast to most other fur seals which wean their young before or at 12 months of age, fur seals in Peru are weaned from between 12 and 36 months. While on land, females move to and from the waterline to thermoregulate . This results in absence of a defined harem structure because females traverse through the territories of several males each day. Daily movements to and from the waters' edge by females carrying young pups, combined with high female densities ashore, resulted in deaths of 40% of pups born throughout the breeding season in 19-84 and 1985. Pup weights and survival to one year of age varies as a function of prey availability at sea . When adult anchovetas are scarce, pup weights and survival are reduced. lnterannual differences in food availability and pup growth rates may be related also to variations in weaning age. The extended and variable duration of lactation of fur seals in Peru may be an adaptive response to depending on a marine environment that fluctuates unpredictably. ACKNOWLEDGEMENTS Throughout the time I took to complete the field work and write up this dissertation, the number of people which in one way or another helped getting it done is enormous! So much so that I cannot possibly mention everyone here. This, of course, doesn't mean I am not grateful for their help. I am. Without any of them, I would have never done it. Thanks! To start, I would like to thank Wildlife Conservation International (a divison of the New York Zoological Society) which almost completely covered the costs of my field research and has now provided me with funds to continue this study for two more years. The Max Planck lnstitut fur Verhaltensphysiologie (Abt. Wickler) , the Percy Sladen Memorial Fund, and King's College, Cambridge also contributed to funding the field work. Bristol-Meyers Co. Ltd. (U.K. & U.S.) provided me with free hair dyes (Clairol-Born- Blonde) to mark the seals. The British Council gave me the initial scholarship to come to Britain and covered 12 months of my time in Cambridge. King's College, the Company of Biologists and my family supported me through the extra-time I took to write-up. While at Punta San Juan logistic support was generously provided by Hierro- Peru, by Pesca-Peru through Dr. Demostenes Cabrera and by the Peruvian Navy through Tavo Salcedo and Gino DiPhilippi. In Lima Ors. Rogelio Villanueva and Abelardo Vildoso helped accelerating and sorting many bureaucratic obstacles at the Ministry of Fisheries and IMARPE for me. Ors. Wolf Arntz at PROCOPA, Daniel Pauly at ICLARM and Cptn. Hector Soldi at the Oceanography and Hydrography Division of the Peruvian Navy, kindly allowed use of unpublished fisheries and oceanographical data. Pamela Saunders, Ken Darcy and Pedro Vasquez collected tag resights and pup weight data for me while I was away in Cambridge. I would have never got involved in fur seal research to the point of spending now almost nine years of my life studying them without the influence of Fritz Trillmich. In 1979, he initiated me into the world of fur seals and since then has supplied friendship and support at all levels. His work on the Galapagos fur seal has provided an invaluable insight into my study population and his comments on my work have kept me from getting lost in some of my wild ideas .. . Fritz and Gerry Kooyman came to Peru with me in 1983 at the beginning of my Ph.D research. Together, they taught me the basics of how to become a fur seal "expert" and then left me to deal with the worst El Nino in the last 400 years by myself . Throughout 1983, Marines Sanchez-Grinan offered great company and 'heroical' help with the seat's analysis . Gerry Kooyman provided a wonderful trip to McMurdo Sound, Antarctica, after I was forced out of San Juan during the guano harvest in September 1983. Gerry has been a continuous source of inspiration as a scientist. His friendship and constant encouragement more than once kept me from giving up on the Ph .D. A great part of the success of this study I owe to Miguelito Goebel who discovered the trick to capture adult female fur seals and their pups and helped in all aspects of the study in 1984, 1985 and 1986. He also provided a highlight to life in the field by producing ~ chocolate truffles! Through Mike I was able to get three of the four dive records described here. Gerry Kooyman and Bob Delong were the generous suppliers of TDRs. The National Marine Mammal Lab. (NWAFC, NOAA) in Seattle provided facilities for the analysis of these records. Dany Gaterio, Jorge Pejoves, Johanne Ouellette and Frances Weick assisted in collecting data in the field. I thank them for their patience in coping with my moods and for their cheerful company. Pedrito Llerena helped in capturing seals, cleaning and storing gear, in keeping an eye on the seals while I was away and in thousand other details for which I am most grateful. While in Cambridge, I have benefitted from discussions with lain 'Gordo', Steve Alban, Robin Best, Seamus Mccann, Phyllis Lee, Rob Harcourt, and Montse Gomendio. All of them also provided very heplful comments on various aspects of the analysis and early drafts of chapters. Rob Harcourt and Nigel Leader-Williams helped proof reading the final draft. To all the members of LARG thanks for their company, support and many useful comments. Many thanks also to the British Antarctic Survey and the Sea Mammal research Unit which provided many facilities and advice, in particular to Christine Phillips, for her help in getting obscure references and allowing me to raid her library. I am also very grateful to the Department of Zoology in Cambridge and all those there who helped in many aspects throughout my time in Britain. John Johnson saved my life by rescuing equipment stranded in Heathrow. Dennis Unwin and Ray Symonds at the workshop invented for me many wonderful measuring instruments. John Andrews and Helen Hewitt provided letters to get me and my gear through the many countries in my path. Neil Maskell and Frances Pang kindly printed the photographs and some of the figures in the text. I specially thank John Croxall and Sheila Anderson for their advice and careful reading of the whole thesis. Without their patience and encouragement I would have never done it. To Tim Glutton-Brock I am most grateful for his supervision. Through his influence my thoughts and writting acquired some logic and learnt the most valuable skill of how to write successful grant proposals! I would also like to thank my whole family. Although they still don't understand what I get from looking at seals, they have never stopped backing me in everything I ever attempted. I am particularly grateful to my fat her and uncles Salvador and Miguel who generously refilled my bank account when it came dangerously close to being empty. My uncle Miguel also allowed me to get the Macintosh computer in which I wrote the entire text and drew most of the figures of the thesis. Finally, I thank all my friends who stood by me through my depressions, provided many good times and made writing up and the British weather bearable ... CONTENTS: Page : Chapter 1 Introduction 1 Chapter 2 Methods and study site 6 Chapter 3 The environment 14 Chapter 4 . Diet 22 Chapter 5 Female foraging behaviour 30 Chapter 6 Breeding behaviour and annual cycle 38 Chapter 7 Attendance behaviour of females 59 Chapter 8 Lactation 71 Chapter 9 Pup mortality 81 Chapter 10 Pup growth 94 Chapter 11 Weaning 101 Chapter 12 Discussion 108 References 1 CHAPTER 1 - INTRODUCTION: Like all other members of the Order Pinnipedia, fur seals (Family Otariidae, Genera Arctocephalus and Callorhinus) breed on land and forage at sea. To forage in a marine environment, seals have evolved a series of adaptations to swim and dive efficiently. These include a hydrodynamic body shape, modification of limbs into flippers , increased oxygen stores and physiological mechanisms which allow efficient use of oxygen reserves . They have also acquired a thick blubber layer and in the case of fur seals a fur coat to reduce heat loss while at sea. Seals show two distinct lactation strategies. Phocids (earless seals) have short lactations (4-63 d), their pups grow fast (0.2-6 kg/d) and in most species females fast throughout lactation (reviewed in Kovacs and Lavigne 1986) . Otariids (fur seals and sea lions) have longer lactations (3-36 months), pups grow slower (0.05-0.1 kg/d) and feed intermittingly while suckling their young (Gentry et al .1986a) . These two strategies make phocids less vulnerable to fluctuations in food availability than otariids. Because phocids depend on stored fat reserves to feed their young, they are more flexible in their timing and location of breeding (Costa et al. 1986). In contrast, otariid mothers have to replenish their fat reserves regularly in order to raise their young and thus depend on a constant food supply to be able to breed successfully . Adaptations for marine foraging may limit the seals' capacity to cope with high air temperatures. Mechanisms to prevent hypothermia while foraging at sea (low surface/volume ratio, increased insulation, high basal metabolic rate) reduce the seals' ability to thermoregulate while breeding on land (Umberger et.al 1986). Thus, seals are most typically found in high latitude areas with temperate to sub- polar climates (King 1983) . However, the South American fur seal (Arctocephalus australis, UNIVERSITY LIBRARY CAMBRIDGE 2 Zimmermann 1783) is found in a diverse range of habitats . It is found all around southern South America, from sub-tropical Peru , around Cape Horn and the Falkland Islands, and up to southern Brazil , covering a latitudinal range of about 35° (Majluf & Trillmich 1981) . Little is known about the ecology of South American fur seals and what limited information is available comes mainly from studies of the population in temperate Uruguay (Vaz-Ferreyra 1956 1971 1982, Ximenez et al. 1984, Ponce de Leon 1984) . The Peruvian population is of particular interest because only it and the Galapagos fur seal (A.galapagoensis) breed on or close to the Equator (Bonner 1981) . The presence of the fur seals in Peru is linked to the flow of the cold Peruvian (Humboldt) current along the west coast of South America (Majluf & Trillmich 1981) . This current dramatically affects the climate and productivity of coastal Peru by reducing air and sea surface temperatures and giving rise to the richest upwell ing system in the world (Idyll 1973, Cushing 1982) . This system sustains a large biomass of pelagic fish comprised almost exclusively of a single species, the Peruvian anchovy or anchoveta ( Engraulis ringens, Fig.1.1 ), that in turn supports large populations of seabirds and marine mammals (Idyll 1973). By breeding in Peru , however, fur seals face at least two major problems. First, their food supply may vary unpredictably from year to year . At variable intervals of between 2-7 years, the upwelling system is affected by major oceanic perturbations known as El Nino (Cane 1983) . These affect the system by increasing sea surface temperatures, reducing primary productivity and directly affecting the depth distribution and abundance of anchoveta (Fig . 1.2) . The effects of El Nino on the Peruvian seabird populations and fisheries have already been well documented (Glantz & Thompson 1981 , Cushing 1982, Arntz et.al 1985), but less is known about how fur s.eals in Peru are affected by these changes (Limberger et al. 1983, Majluf 1985, Trillmich et al. 1986) . Second , although air temperatures in Peru are lower for the latitude range than the expected due to the effects of the Humboldt c.urrent (Hanwell \ 80 ' 1. \ : I ' ' ' • VALP~RAISO ' I ,- ~"' --. 1. 1 The Peruvian upwelling system. Anchovetas normally occur in a northern and southern population (shaded area) (from Idyll 1973). x_ ~ ~ "O ' E. (J) -"-0 - .:::~ > C, '- OI :, 0 (f)- Period of relatively stable prerecruit survival, and of relatively high anchoveta stocks El Nino ~ Overfishing effect: Major drops in prerecruit survival not associated with El Nino events Period of low anchoveta stock and of highly variable _pre recruit survival El Nino EI Nino A. r--~---, ~ 1962 . 1963 1964 . 1965 1966 1967 1968 1969 1970 . 1971 1972 1973 . 1974 . 1975 1976 1977 1978 1.2 Survival of anchoveta eggs and larvae. This index reflects variations in the stocks of adult anchoveta (from Pauly and Tsukayama 1987). 1980), intense solar radiation at low latitudes increases the effective temperature of the environment as perceived by the animal (Umberger et al. 1986) and thus may cause problems of overheating to the fur seals breeding and suckling their young ashore . One of the most striking characteristics of the behaviour of the S.American fur seals in Peru (hereafter referred to as Peruvian fur seals) is that some females suckle their young for very long periods . Females often suckle their young until they are two years of age, or longer (Majluf 1987), with the result that often two, and sometimes up to three, offspring of different ages suckle simultaneously (Fig . 1.3). In contrast, most other fur seals always wean their young before they are one year old (Bonner 1984) . Furthermore, in the populations of the S.American fur seal in Uruguay (Ponce de Leon 1983) and the Falkland Is. (Bonner 1984), females rarely suckle yearlings. Only Galapagos fur seals regularly suckle their young for more than one year (Trillmich 1984) and this species is also affected by El Nil'lo events (Trillmich & Umberger 1985) . In the Galapagos seals, suckling a yearling or older offspring increases the probability of mortality of the mother's subsequent pup and reduces her fertility (Trillmich 1986) . Are fur seals in Peru similarly affected by having their young suckle for more than one year? If this is so, why don't females wean their young earlier? Are long lactations related to living in the tropics and/or with having a fluctuating food supply? In this study, I examine the adaptations of the Peruvian fur seals to their environment by looking at their breeding ecology and behaviour. In particular, I examine how changes in food availability and breeding in a hot environment affected the duration of lactation and the females' breeding success . This was possible in a short-term study ( 1983-1986), because conditions at sea varied widely . The project started in 1983, during the most intense El Nil'lo in the last 400 years (Quinn et al. 1986), and in 1986/7 there was a second, milder event. In 1 . 3 A female South American fur seal at Punta San Juan simultaneously suckling a newborn pup, a yearling and a two year old immature. 4 1984 and 1985, conditions at sea were apparently good (Chap.3) . The data presented in the chapters to follow were mainly obtained in the two good years of 1984/5 and 1985/6. Only limited information was obtained during the severe 1982/3 El Nif'lo but where relevant , data from this year are included to show how fur seals were affected in conditions of extreme food shortage . One must bear in mind however, that the 1982/3 El Nif'lo was exceptionally strong, thus, its effects are likely to be considerably greater than events of regular intensity. I concentrate almost exclusively on the ecology and behaviour of the adult female fur seals and their young. However, because little is known about the species in general, I also briefly describe the breeding behaviour of males. The study was carried out at Punta San Juan (Fig . 1.4). where more than 50% of the Peruvian fur seals breed (Majluf & Trillmich 1981) . The study site and methods are described in Chapter 2. Thorughout the thesis, I place special emphasis on anchoveta because of the importance of this resource to the Peruvian vertebrate marine fauna . Until 1973, it comprised up to 95% of the total pelagic fish biomass off Peru (Idyll 1973) . Since fur seals are mainly pelagic , oportunistic -feeders (Kajimura 1984, Gentry & Kooyman 1986). they should be affected by changes in the availability of this species. The thesis is divided into two parts . In chapters 3-5 the feeding ecology of lacta1ing females is examined . First , I describe the diet of fur seals from the identification of hard parts (fish otoliths and squid beaks) in fresh faecal material (Chapter 4) . Second, I look at the diving behaviour of adult females and compare their dive patterns to the known daily, seasonal and annual changes in vertical distribution of anchoveta obtained from fisheries data (Chapter 5) . Both, diet and diving patterns are examined in relation to the seasonal and yearly variations in primary and secondary productivity at sea which are described in Chapter 3. The second part of the thesis (Chapters 6-11) deals with the breeding I 5 • 'Ecuador· c.;.J . . I ~. ...... . ~ ( ....,, . I O 0 I 5 • 8 0' 7 5 ° PERU i ma r I ( ,...) I' r / BRASIL \ '· \ \ /...._ '\ I l Norlhgrn bretdlng Ii mi I I 5 ' 7 S 0 1. 4 Coastline of Peru with location of Punta San Juan and northern breeding limit of South American fur seal. 5 behaviour and factors affecting weaning and reproductive success in the Peruvian fur seals . In Chapter 6 I look at the annual cycle and basic breeding patterns of male and female fur seals at Punta San Juan. In Chapter 7 I examine how females divide their time between foraging at sea and suckling their young on land (attendance patterns). In Chapter 8, by looking at the sucking behaviour of pups and yearlings, the amount of time spent by females suckling young of differing age or weight while on land is calculated and used as an index of the energy demands of the young. The factors affecting pup mortality are examined in Chapter 9. Here I examine how high temperatures and/or changes in food availability affect pup mortality in the fur seals at Punta San Juan. The seasonal and annual variations in pup growth are examined in Chapter 10. Of the young that survive to the next breeding season, how many continue to suck for a second year? Does suckling a yearling affect its mother's fertility? These questions are examined in Chapter 11 by following tagged individuals from one year to the next. 6 CHAPTER 2 - METHODS AND STUDY SITE In this chapter I describe those methods which are common to several of the chapters to follow and give a brief general description of Punta San Juan, the study site . When the methods concern only a limited section of the data, a detailed description is given in the relevant chapter. Also , the methods here given were used mainly in the 1984/5 and 1985/6 seasons . Throughout 1983 various methods were tested and some discarded. Few of the data collected in 1983 were used here and methods specific to these are noted where appropriate . Seasons studied: The results discussed below were obtained between January 1983 and January 1986, covering two complete breeding seasons (October - December) and part of the non-breeding periods in between. Details of the dates covered in this study are given in Table 2.1. Additional information on tag resights and pup weights obtained at times when I was away from Punta San Juan were provided by Pamela Saunders and Ken Darcy (Dept . of Zoology, Cambridge) for July- August 1985 and by Pedro Vasquez (Depto . de Manejo Foresta!, Universidad Nacional Agraria, Lima) for February-March, August and December 1986 and January-March 1987. When data from these sources are used, this is so noted in the text . The study site: The study was carried out at Punta San Juan, which is the breeding site of almost 50% of the S. American fur sea! population in Peru (Majluf & Trillmich 1981 ). Punta San Juan is a peninsula cut ott from the hin1erland by a concrete wall (Fig . 2.1 ). Fur seals breed on most of the beaches which are distributed amund the peninsula. Most of the observations and data here described wern obtained from beach S3 , where the highest concentratfon of fur sBals at Punta Table 2.1- Dates during which data were collected in this study Year 1983 1984/5 1985/6 1986 Dates Jan . 10 - Feb. 02 Mar. 03 - Apr. 03 Apr. 12 - May 09 May 21 - Jun. 01 Jun. 08 - Sep. 11 Oct. 03 - Dec. 12 Jan. 11 - Feb. 22 Sep. 09 - Dec. 28 Jan. 06 - Jan. 16 Nov .08 - Nov. 15 15°22' 2.1 LOU 0 1100 me le r s 1: 10,000 " 1000 : cone rete . well Punta San Juan. Shaded areas indicate the different beaches where fur seals and sea lions are found. 7 San Juan is found. This beach has an area of approximately 750 m2 and is surrounded by a 20-25 m high cliff (see Chapter 3 for more details). Access to this beach is limited to one narrow path which was constantly visible from the main observation point from which data were recorded and thus , human disturbances were kept to a minimum. From 1984, all watches were carried out from a blind (A) which gave a view of the entire length of the study area. In 1985 a second blind (B) was built to cover the remainder of the beach. The position of these blinds is shown in Fig. 2.2. To enable comparisons with other areas, some data were also collected from beaches N5 and S4 (see Fig. 2.1 ). Climatic and oceanographic variables: From October 1984, environmental data were recorded three times a day (06:00 , 12:00 and 18:00) . Most measurements were taken from blinds A or B, except for the sea surface temperature (SST) which was recorded from the cove in front of the houses, on beach N1 . Wind direction was recorded from a wind vane put on top of the main house. In 1985/6 ground , air and sea surface temperatures were measured using a hand-held digital thermometer (range -50 to 150 °c: Edale Instruments (Cambridge) Ltd. ) with interchangeable thermistor probes at the end of PVC covered leads of varying lengths. Ground temperature was measured at two different points immediately below each of the blinds using 60m long ieads. One of the probes was buried five cm under the ground and the other was placed on the ground where it was shaded and held firm by a heavy rock . Air temperature was measured using a short lead (1 . 5 m long) permanently fixed 20 cm under the roof in blind B, where it was shaded and well ventilated. To record SSTs, a probe on a seven m long lead attached to a three m long bamboo pole was used. This probe was thrown out to sea as far as the cable would reach and the temperature measured at this point. In 1984/5 a "Portable Weather Station" de-signed by R. J. Symmonds and D. M. Unwin (Electronics Section, Dept. of Zoology , C) ---a------------ -- -- 0-~ ---. __ • _J""--, -:~ o.O • • c;:, 'O V e"-U ·------co ~-..=-..:::· o o.,, 0 ° ()d··:. ~ •'-or"" ~ - - ~ , ~25 m contour surrounding the beach (vertical shading), observation blinds (A & B) and census areas (dotted lines). 8 Cambridge), was used. This instrument combines a thermistor temperature probe (0-50 °c), a thermistor anemometer (0 . 05-20 ms-1) and a silicone photo- diode to measure radiation (0-1 KWm-2) , again using exchangeable probes. Ground and air temperature were measured in a similar way as above (only from blind A in this case) . Both instruments were calibrated at the beginning of the 1985/6 season. There was no significant difference between air temperature values, but their values for ground temperature varied widely. For the latter, only the 1985/6 values are here used. SSTs in 1984/5 were taken using a mercury thermometer submerged in a deep, well mixed tide pool on beach N1 . In both seasons a max-min thermometer was placed inside blind A. Every day it was reset at 06 :00 and measurements taken at 18 :00 . Black bulb temperature (BBT) was recorded from a black bulb globe thermometer (range 0- 65 °c; Casella & Co. , London) which hung from a metal rod permanently fixed to one of the sides of blind A, where it received direct sunlight throughout the day. Cloud cover, sea state and swell were estimated by eye by the person recording the data at the time. Cloud cover was originally given in Oktas, but here days have been classified as clear, if the sky was less than 50% covered, and overcast, if clouds covered more than 50% of the sky. Sea state was classified according to the folowing categories : dead calm, calm, choppy, white horses, many white horses, rough and very rough. Swell was classified as : none, slight, moderate, heavy and very heavy. Censuses: Throughout the 1984/5 and 1985/6 seasons, every day between 06 :00 and 07 :30, a portion of beach S3 (covering approximately 50% of the study area) was censused. These censuses were made from blind A with 1 OX binoculars and recorded on specially designed check sheets which were divided into sections corresponding to the different quadrants painted on the beach (Fig . 2.2) . These quadrants varied in shape and size and were defined either by a change in substrate or by painted rocks ; in general they were just used as an aid for 9 censusing. For each census the following age and sex groups were distinguished: adult (territorial) and sub-adult males, adult females , pups, yearlings, older immatures and indeterminates (this last usually included only young females or sub adult males which were difficult to distinguish) . At the same time, the number of adult females observed suckling a pup, yearling or both simultaneously were noted. Every three or four days, these censuses were carried out every three hours through the day. During the 1985/6 season, the position of individually known territorial bulls was recorded at the end of the morning census. Simultaneously with the above censuses , a second observer walking daily along the cliff surrounding beach S3 counted the total number of live and freshly dead pups, as well as the number of fresh placentas in the entire study area . Generally, pups dead for more than one day were either obviously bloated or opened by turkey vultures (Cathartes aura) ; thus, double counting of dead pups was avoided. Because the whole study area was well protected from the swell by the intertidal rocks, pup carcasses tended to remain ashore until totally decomposed and only rarely were seas strong enough to wash the beach clean of carcasses. Therefore, the numbers of freshly dead pups are assumed to reflect pup mortality quite accurately. Similar but slightly less detailed censuses were carried out daily and at the same times as above on beach N5 by a third observer (only in 1985/6) . In this case, yearlings and 2-3 year-old animals were counted as immatures, and adult and sub-adult males were counted as males. The same person also obtained numbers of bachelor males on beach N4, which was not used for breeding by fur seals . Capture and restraint: Only adult females, pups and a few yearlings were handled in this study . Capture methods varied according to the breeding status and age of the animals 10 involved. Adult pregnant females were only captured early during the 1984 breeding season . These were immobilized with a mixture of Ketamine hydrochloride ( Vetalar, Parke Davis/Warner Lambert , Usk Road, Pontypool , Gwent, NP4 8YH, U. K.) and Valium (= Diazepam, Roche Products Ltd. U. K. ). The latter came in a 10 mg/2ml solution in which the Ketamine powder was dissolved to keep the total volume at two ml. A maximum dose of 400 mg (6-7 gkg-1) of Ketamine was applied per animal (mean female weight= 60 kg, see Chapter 6). To inject the drugs, a dart-gun (lnjekta Jr. ) which shoots two ml syringes propelled by compressed air was used from a distance of approximately 20-30 m. Ten minutes after the drug was applied, the animal was stalked and captured using a 230 or 183 cm circumference hoop-net (Fuhrman Diversified) . The effects of the drug usually wore off 30 min after darting but the animal was kept under constant observation for at least one hour, in order to make sure she did not go to sea while under the effects of the anaesthetic . Later on in the season as more females gave birth, only females with very recently born pups were captured. These were not immobilized. Instead, the strong attachment of females to their pups, which was typical of the perinatal attendance period, was exploited. To force the females to follow us, their pups were taken away using five m long bamboo chokers (poles with a loop of rope at the end, Gentry & Holt 1982). Once well away from the territorial bulls and as far back in the beach as possible, the females were captured either with a hoop-net or a heavier wood choker. In this way, both mother and pup were captured at once . After capture, adult f-emales were kept immobile using a 'restraint board' (Fig. 2.3), which locked the females' neck in a diamond shap.ed opening while her flippers were kept close to the body with tightly fitting straps. Pups and yearlings were captured either by directly grabbing their hind flippers (only pups) with a short metal or wood choker, or with a small (137 cm 2.3 A female fur seal being weighed on the r-estraint board . Her head is being held bya long wooden choker used for captures. circumference) hoop-net. No restraint board was used for small animals. When possible, large groups of pups were rounded up inside a cave or at the top of a steep slope and then handled one by one . Weights and marking: Adult animals were weighed using two 50 Kg PESOLA spring balances suspended from an aluminium tripod with telescopic legs . Depending on whether the female was captured with a choker or with a hoop-net, she was then weighed while held either on the restraint-board or inside the net (Fig. 2.3). Pups and yearlings were weighed inside a nylon mesh bag, using either a 20 kg PESOLA spring balance or one of the 50 kg balances used for the adults. When only a few animals were weighed, the weighing gear was lifted by hand. When large numbers of animals were being processed, the aluminum tripod was used, only this time without extending the legs to their full length. For some individuals total length, fore flipper length and axillary girth were measured using a metal tape. Various methods of marking were used during this study. In 1983, because I was working by mysett, the animals were only marked from a distance by throwing blown egg shells full of waterproof enamel. This method produced only a very short term mark because animals tended to groom the paint off very quickly. A few animals in 1983 were tagged in both front flippers using plastic numbered tags (Allflex Medium yellow/yellow combination). From 1984 onwards, most animals were tagged using the same type of tags (1984-yellow, 1985- green, 1986-white) in all cases combined with yellow for female pups, red for male pups, green for yearlings and blue for adult females. Table 2.2 lists the number of animals tagged each year. Animals which were studied in more detail were further marked by clipping the hairs on their backs in the shape of large letters, numbers, symbols or words . To make these marks last longer, the underfur left after clipping the guard hairs was bleached using a ladies hair bleach with a peroxide base (Clairol Born Blonde or Clairol Ultra Blue, Fig. 2.4) . These marks 2.4 A marked and tagged fur seal female with pup. The mark was made by clipping the guard hairs and bleaching the underfur with Clairol Born Blonde. Table 2.2 - Numbers of fur seals tagged in this study --------- ---- -- -- ---------- -- ---- -------- ---------- -- -- ---------------- -- -- ------------ ---------------- -------·-- Year Ad.Jk Females Males Pu~ Yearlirgs Females ----------------------------------------------------------------------------------------------------------------- 1983 1984/5 1985/6 1986 3 64 22 50 11 202 209 127 6 190 174 123 3 10 3 ------------------------------------------------------------------------------------------------------------------ Total: 139 549 493 16 12 lasted at least six months for the adults and yearlings, and 3-4 months for the pups (until they moulted) . Some tags remained on for more than three years, but in many cases, the numbers have been erased by the constant rubbing against the rocks (in particular the 1983/4 yellow tags) . In general, the adult females' tags tended to stay on and remain visible for longer than the pups' . Resightability varied acco-rding to how the individual was marked. Animals which were both clipped and tagged were the most visible , followed in decreasing order of resightability by animals which were only clipped, tagged only and paint marked. A few territorial males were marked by dabbing them with brightly coloured waterproof enamel applied on a sponge at the end of a long (>5 m) bamboo pole. These marks were visible for at least one month, but this seemed to depend on whether the animal was wet or dry when marked. For longer term recognition, old scars on their mane or flipper were used. Tag checks: Every day at 07:00, 12:00 and 17:00 (GMT-5), beaches S3 and N5 were scanned for at least an hour to see which tagged/marked animals were present. During these checks, for each animal resighted it was noted whether ~ was alone or with its mother (or offspring, if it was an adult female), if more than one young were with the female simultaneously, if it was suckl ing or not , or if it was just arriving or departing. These notes were then transcribed onto individual cards for each animal. In 1984 tags were read with 1 Ox40 binoculars and notes kept in a notebook. After 1985, a 30x75 telescope was used and notes taken with a microcassette portable recorder and later transcribed. Statistical analysis and graphics: Most data analysis was carried out using either the Cambridge University IBM mainframe computer or a Macintosh Plus micro computer. Statistical analysis was carried out using either SPSS-X (SPSS Inc. ), Statworks (Heyden & Son) or Statview 512+ (Brainworks Inc. ) . Most of the graphics were made using either 13 Statview 512+, Cricket Graph (Heyden & Son) or Microsoft Chart. 14 CHAPTER 3 - THE ENVIRONMENT In this chapter I describe the marine and terrestrial environment of the fur seals at Punta San Juan. In the first part , I briefly examine the major physical and biological characteristics of the Peruvian upwelling ecosystem and its short and long term variations. Because abundant information on these topics has been recently published elsewhere (Glantz & Thompson 1981 , Arntz et al. 1985, Robinson & Del Pino 1985). I treat only those parameters, which affect the marine environment of the fur seals . Thus, I examine first how changes in sea surface temperatures (SST). and primary and secondary productivity at sea affect the behaviour, distribution and abundance of the fur seals' potential prey items, in particular, pelagic fish such as anchoveta. In the second part, I describe the climatological conditions that the seals have to endure while on land during and immediately after the breeding season at Punta San Juan. Here I concentrate mainly on the thermal environment since it is the main parameter constraining the fur seals' behaviour on land (Umberger et al. 1986). THE MARINE ENVIRONMENT Physical factors Coastal upwelling is the process by which subsurface water is brought to the surface and moved away from the area of vertical transport by horizontal surface flow (Fig . 3.1) . As the water from the surface is swept northwards by the local prevailing winds (SE) and westwards by the Coriolis force, the colder subsurface water slowly rises to replace it. This deeper water is cool and rich in inorganic plant nutrients such as nitrates, phosphates and silicates, but is separated from the surface layer by a thermal gradient (thermocline) . Water upwelling to the surface comes from just below the thermocline . This supply of PERUVIAN COAST OCEANIC CURREN T 3. 1 Vertical and horizontal flow components of the Peruvian upwelling system (from Idyll 1973) . 15 nutrients to the surface layer is unusual for warm tropical waters, which are typically limited in nutrients . This enrichment enhances primary productivity by phytoplankton, and the abundant organic matter produced is passed to zooplankton, fish, birds and marine mammals through grazing and predation, (Barber and Smith 1981). These improved conditions continue as long as winds are favourable to upwelling, which is generally throughout the year. Thus, the annual primary production in upwelling areas is much higher than it is in temperate or high latitude ecosystems (Barber & Chavez 1986) . The coastal region off the west coast of South America is considered to be the most productive in the world (Ryther 1969, Idyll 1973). The depth of the thermocline, which determines the availability of nutrients to the system, is directly dependent on the persistence of the easterly trade winds along the equator (Cane 1983) . When these blow westwards across the Pacific, warm water accumulates in the west (Indonesia) and this originates a thermocline slope which brings the thermocline closer to the surf ace off the west coast of South America (Fig . 3.2a) . For upwelling to be able to reach the nutrients, the thermocline has to be at a depth within reach of the upwelling process (40-80m) . Thus, by creating the east/west thermocline slope, winds control the nutrient (and temperature) content of the water entrained into the upwelling circulation as well as providing the local driving force for vertical transport (Barber & Chavez 1986). Seasonal variations In the Peruvian upwelling system, there is a clear seasonal variation in the intensity of upwelling . It is most intense during the winter (June -October), and lowest in the summer (Zuta et al. 1978, Vinogradov 1981) . There is a relationship between wind speed and SST (inverse) with intensity of upwelling . In the summer (January - March), when SSTs are highest and the local winds are weak, upwelling is very low, whereas in the winter when SSTs are lowest and local winds A TRADES --) SEA LEVEL - -A - - - --- I --1 ..- ~ 1£om oc\.'~ _.),- / ·n-\t~ - ;s. AMERICA -- I < :x --/\ r - 1, -4-- j I B TRADES r::: - SEA LEVEL/ -- ----r- f -- ) II 100m _ 1/ _ ___..t_ \/ I 3.2 Differences in depth of the thermocline along the coast of South America, in relation to intensity of Equatorial trade winds, sea level and direction of vertical flow (thin arrows) during (A) stable and (8) El Nino conditions. 16 are strongest, upwelling is most intense (Zuta et al. 1978). During the winter, an influx of warm oceanic water interacting with the cold outflowing upwelled water, results in a meander-like shape of the isotherms (Zuta & Urquiza 1972). This influx of oceanic water is related to a relaxation of the southeast trades along the equator and appears to result in a winter deepening of the thermocline (Myers 1979, Cane 1983, Chavez 1987). In the summer, the isotherms are parallel and close to the coast (Zuta et al. 1978), the Humboldt current is at its narrowest (Idyll 1973) and the thermocline shallower (Myers 1979). The source of water entrained into the surface layer also changes with season: in the summer upwelling brings waters of subantarctic origin to the surface while in the winter the sources are mostly from the north (Zuta et al. 1978). Long tenn variations: El Nil'\o Changes in the large scale pattern of wind circulation will directly affect the productivity of the system (Cane 1983, Barber & Chavez 1986, Ramage 1986). At irregular intervals of several years, the Equatorial Trade winds relax and initiate an unstable condition known as El Nil'lo (Fig. 3.2b) . As the trades weaken, the water which has accumulated in the west Pacific forms a gig-antic Kelvin wave and, in less than two months, crosses the Pacific, causing a massive influx of warm water along the South American coasts. The east-west th-ermocline slope that exists under normal trade wind conditions disappears, and as more and more warm water invades the system, the thermocline is progressively deprnssed and nutrients become unavailable to the upwelling process (C·ane 1983, Barber & Chave-z 1986). Coastal upwelling continues, b_ut the water transported to the surface is warm and poor in nutrients (Cane 1983, Kerr 1983, Fonseca 1983=). This reduction in the nutrient supply causes primary productivity to decreas-e and eventually affects all the higher trophic leve-ls in the system. (Arntz 1986, Barber & Chavez 1986) . Depending on the magnitude of the ev-ent, normal upw-elling conditions might persist in the southern portion of Peru and Chile. Persistent 17 patches of cool waters close to shore have been described, in particular around 15 os (Villanueva et al. 1969). Similar events on a much smaller scale take place every year around late December (hence the name El Nit'lo : the Child) but these have only local effects limited to the Galapagos, Ecuador and northern Peru (down to 5 OS) . The anomalies described above happen at the same time of the year but usually affect a much greater portion of the South American coast (at least down to 12 OS) and only occur at irregular intervals ranging between 2-1 O years and on average every four years (Cane 1983) . At present only the large scale events have kept the name El Nit'lo (Quinn & Neal 1982, Cane 1983). Biological factors The Peruvian upwelling ecosystem sustains a large flow of living matter. Within a narrow band covering only a tiny fraction of the worlds' ocean surface it was able to supply up to 22% of all the fish caught in the world. In the Peruvian system most of the energy flow stops with the anchoveta, as opposed to most other systems where primary productivity is progressively utilized by complex food chains (Idyll 1973) . Up to 1973, anchovetas comprised 95% of the total pelagic fish biomass off Peru (Quinn & Neal 1982) . Since then, the anchoveta population has declined substantially and fluctuated markedly (Cushing 1982, Pauly and Tsukayama 1983, see Fig. 1.2). Anchovetas are apparently highly adapted for life in the Peruv.ian coastal upwelling system (Barber & Chavez 1986) . Because of their strong preference for cold waters (16-18 °c, Cowles et al. 1977), their movements are not random, but vary according to changes in the shap.e of the Peru current and with SSTs (Idyll 1973) . This behaviour normally leads them to the areas where plankton is most abundant (Barber & Chavez l986). During the summer (January - March) , when the coastal current is at its narrowest, they are densely concentrated in sha~low waters and close to the shore . At this time of the year they are readily UNIVERSITY LIBRA!W CAMDRiDC~ 18 available to the fisheries (Fig. 3.3), to the seabirds and to the seals (Idyll 1973, Arntz 1986) . During the winter, when warm oceanic waters invade the system, they disperse and remain at greater depths, below the now deeper thermocline (Jordan & Chirinos de Vildoso 1965, Saetersdal et al. 1965) . Spawning occurs in August and September during the winter and again on a smaller scale in January and February (Jordan & Chirinos de Vildoso 1965, Idyll 1973) . Anchovetas are known to carry out daily vertical migrations. During the day, the schools remain in deep water, concentrated in tight patches (Fig. 3.4) . Around dusk, they start dispersing and migrating upwards, until they reach the upper layers (<25 m) . By then, they are dispersed and remain so until dawn, when the whole process is reversed (IMARPE, 1969) . During spawning these migrations might not occur (Jordan, 1971). When El Nit'lo events take place, the decrease in primary productivity and the elevated temperatures affect the anchoveta's distribution, growth, survival and reproductive success . During the onset of El Nino , seeking out cooler waters causes them to concentrate in the persistent inshore upwelling . This behaviour is adaptive if the event is weak, but during strong El Nino events, these cooler patches become traps for the anchoveta. As the event progresses, they are surrounded on three sides by warmer water and eventually, the cold patches disappear and the Hsh perish. The fish that do survive, are emaciated and although they might still manage to spawn, larval survival is severely reduced (Barber & Chavez 1986) . The surviving adults tend to remain at greater depth than normal, usually below the thffrmocline which is now depressed and the vertical migrations of the schools are restricted to deeper layers or stop altogether (Arntz 1986) . Thus , under El Nino conditions, anchovetas become less abundant and virtually inaccessible to the seais, birds and fisheries (Cushing 1982, Arntz et al. 1985, Quinn & Neal 1982) . Punta San Juan: l983 - 1987 >< UJ Cl z >- ~ ...J Q'.] 0.4 0.3 0.2 0 .1 J M ' '· .. -..... ,'q• ' .. ',• . .. ·,' . \ ·. '·. 'u' \ ', \ ·. Ttmpt1oluro M ,·. ' \ ., ~ ·. \ ·. \ · .. \ ·. \ \\.: 'o' , ... J MONTH .... ·· I I I ,.o I O O O O O O O I o,,, ~ ,' , • ""· s / :->I -c I I ' S· C .•. ,,.·· l l9'C 10 l 7 lu l~ N 3. 3 Seasonal variation in availability of anchoveta (calculated from catch per unit effort data from the Peruvian fishery) and sea surface temperature (SST 0 c) for northern (N), central (C) and southern (S) Peru . (from Jordan 1971 ). DAY DUSK 3 NIGHT \ :.;_._ ., ... ~ V • 3.4 Vertical migration patterns of anchoveta from echosound traces (IMARPE 1969). Darker patches indicate high concentrations of anchov~etas. Depth shown in meters. Dark band at the top shows shade appearing on the original photograph. 19 Punta San Juan is the area of coldest waters , most intense upwelling (Zuta et al. 1978) and highest primary productivity (Cushing 1982) off the coast of Peru . Because of these factors, combined with its southerly location (15 os) and the very intense local winds and extremely narrow continental shelf, upwelling tends to persist for longer around this area even when surrounding areas are affected by El Nit'lo conditions (Villanueva et al. 1969) . Fig . 3.5 shows the monthly variation of SST and wind speed at Punta San Juan. The yearly SST variation around the mean at Punta San Juan between 1958 and 1987 is shown in Fig . 3.6. In this figure , all occurrences of El Nit'lo appear as positive anomalies, but not all positive anomalies are considered El Nit'lo events (based on the list of "officially" recognized El Nit'lo events , Quinn · 1978). During this study, there were two El Nit'lo events; one in 1982/3 considered to be the most intense El Nit'lo in the last 400 years and a second, milder event in 1986/7. Fig.3.7gives the mean monthly SST recorded at Punta San Juan throughout the periods covered in this study and shows the anomalies recorded in 1982/83 and 1986/7. In 1982/3 SST rose to 8 °c above the normal temperature for the time of the year, and the thermocline was depressed from the usual Som to about 100m (Fig.3,a . In 1986/7, SSTs rose only to 2°c above the monthly mean , but little is known yet of the effects of this event . 1984 and 1985 are both considered to be non-El Nit'lo years with1985 being exceptionally favourable . THE TERRESTRIAL ENVIRONMENT Physical factors The Peruvian coastline is backed by a long narrow band ( < 80 km wide) of extremely arid desert bounded by the Andes. Punta San Juan is located at the edge of this desert . Here , tur seals inhabit mcky and oft.en quite steep slopes, some of them exposed to heavy surf . The bead=1es are backed by steep cliffs so that for mo-st of the day some places provide shade (Trillmich & Majluf 1981) . I- (.f') (.f') ..... 0 uJ uJ a. (.f') 0 z ): 17 16 •ssr OWIND SPEED 15 14 13 12 11 10 9 J F M A M J J A s 0 N D MONTH 3 .5 Seasonal variation in SST (mean for 1958-1974) and wind speed (knots, mean for 1960-1973) off Punta San Juan (from Zula et al. 1978). ~ w I: >- _J ::c ..... z 0 I: w ::c ..... I: 0 a::: LJ.. Cl) z Q 1- 4: > w Cl I- C/) Cl) 7-1--~--~----~------------~-_....--~----~-- 6 5 4 3 2 0 -1 -2 -3-t------.----.----,-----,---.-----,---.....--...,.....--~---.----.-----,---~-----+-55 60 65 70 75 80 85 YEAR 3.6 Deviations in SST from the monthly mean (1958-1 987) at Punta San Juan. El Nino events are represented by peaks exceeding one S.D. from the mean of the deviations over the entire period (da1a courtesy of Dr.D.Enfield, Oregon State University) . 90 + S.D. µ - S.D. 24 22 20 '"' u 0 18 S../ I- (/) (/) 16 14 12 •02/3 084/5 085/6 ()86/7 s 0 N D J F M MONTH 3. 7 Monthly mean SST at Punta San Juan during the study period (data courtesy of Dr. D.Enfield, Oregon State University) . 0 .,........._ u ~ 100 :i:: I- a. w Cl 200 300 3.8 10 I I I I SEA TEMPERATURE (m) 15 20 25 _L _________ ...__, ______ __,_ _ _ (61. /10)- (71. /175)-..... (75110) / / / { _\_ _1 -/ \ I - - - / ............. - ---- I .,, -::, .,. I ,.,,/"" .,.. .... °'-(73/60) - - "" - - - ............... (72/80) Temperature distribution with water depth off Punta San Juan under normal conditions (solid line) and during El Nino 1982-83 (dashed line) (data courtesy of PROCOPA / IMARPE - Callao). 20 Beach S3, the main study site, is a long, narrow beach (approximately 150 x 50 m, Fig . 2.2). It consists of a slope covered in guano that is protected from the heavy surf by a rock shelf along the extension of the beach. At low tide, this shelf provides tidepools where the seals can keep themselves wet during the hot hours of the day and where pups can swim and play safely away from the heavy surf. Conditions at sea (sea state and swell) were generally moderate . Mean monthly air temperatures varied within a narrow range of between 16 and 22 oc throughout the year. The coolest time of the year was between June and October (Fig . 3.9). Between September and January before, during and immediately after the fur seals' breeding season, air temperatures increased steadily with time, ranging between 15 and 34 oc. Both maximum and minimum air temperatures were significantly higher in 1984 than in 1985 although the diurnal temperature range (max - min) was significantly narrower in 1984 (t64= 2.596, p=0.012, Fig. 3.10) . However, because the dark coloration of seals tends to absorb more heat, a better indication of the heat load to which the seals are exposed is given by the black bulb temperatures (BBTs) . These peaked around noon, when solar radiation was most intense, and on average were about 6 oc higher than the maximum air temperature during the day. This difference was greater in 1985 than in 1984 (XBBT- max = 84/5 : 6.5 oc, 85/6: 10. 9 oc; t33= - 4. 298, p < 0.0001) althol:lgh there was no significant ditterence between BBTs at noon in the two years (XssT = 8-4/5: 35.3 ± 0.47 oc, n = 70; 85/6: 34. 3 ± 0.38 0 c . n = 97) . In both years. about 70% of the days observed were clear(< 50% cloud cover) , though early in the morning (0600) and later in the afternoon (1800) tended to be significantly more overcast than around noon (G = 13. 76, p < 0. 005) . Biological factors At Punta San Juan fur seals oft-en share their breeding beaches with S.American sea lions ( Otaria byronia) and with Humboldt penguiAs ( Spheniscus 23 22 21 ,....._ u 0 '--' LLJ 20 O'. ::) I- < O'. 19 LLJ n. I: LLJ I- 18 ~ < 17 16 15 J F M A M J J A s 0 N D MONTH 3. 9 Seasonal variations in mean monthly air 1empertaures at Punta San Juan (means for1949-1973) (from Zuta et al., 1978). 36 34 ,..... (.) 32 0 '-" (J) 30 LJ.J ~ ::, 28 I- < ~ 26 LJ.J a.. I: 24 LJ.J I- ~ 22 < :?; 20 I: I 18 X < I: 16 14 12 •MAX 84 OMAX 85 •MIN 84 OM IN 85 • •••• ••• • ••••• • • • • •••• •• ••• • • • • •• • 0 • • 0 O 0 • 00 0 00-e • • 0 0 0 Oo • • 0 0 • o. 0 0 • 0 00 • •• 0 0 0 0 "•·· 0 0 .o O oo 0 0 0 0 0 Iii 0 0 o•• • • • • • ••••• • 0 ••••• • • •• D •••• • • • Do OD D 000 OCTOBER NOVEMBER DECEMBER 3.10 Variations in maximum-min air temperatures at Punta San Juan throughout the fur seals' breeding season in 1984 and 1985. Max 84: y = 26.83 + 0.25x, n = 58, R2 = 0.03, n.s. Max 85 : y = 19.63 + 0.74x, n = 98, R2 = 0.716, p = 0.0001 Min 84: y 18.38 + 0.03x, n = 58, R2 = 0.157, p = 0.0015 Min 85: y = 14.54 + 0.45x, n = 98, R2 = 0.769, p = 0.0001 0 21 humboldti1) . Between these three species there is little competition for breeding space because each appears to show a preference for different types of substrate . Fur seals generally breed on protected rocky areas with access to tidepools, sea lions on flat sandy beaches, and penguins in caves, holes or slopes at the back of the fur seal breeding beaches or on top of the cliffs surrounding these areas. Interactions between fur seals and sea lions while ashore are limited to disturbances by sub-adult male sea lions to the fur seal rookeries when they make incursions to capture pups (see Chapter 9 for more details) . Otherwise, fur seals generally ignore the presence of sea lion females or immatures (0-2 years old) . Penguins going across the fur seal rookeries, when coming to and from the water, are also generally ignored by most fur seals. Pups may sometimes try to play with them. Fur seals at Punta San Juan are in general well protected and suffer few disturbances except when there is a guano harvest. These take place every 2-4 years . At these times 300-400 of workers are brought in for periods varying between 3-6 months to extract the layer of guano deposited in the top part of the peninsula. The presence of the workers cause great disturbances to the fur seals because they raid the beaches to capture pups for their fur. The effects of these disturbances have not been quantified. 22 CHAPTER 4 - DIET In this chapter I examine the seasonal and annual variations in the fur seals' diet to determine whether anchoveta is an important item, despite its decline in abundance since 1973, and if so, whether changes in the distribution and abundance of this species due to El Nif'lo affect the fur seals' diet composition . Fur seals are pelagic feeders whose diet cannot be easily documented by direct observation. Indirect methods , such as the identification of hard parts (fish otoliths and squid beaks) in seats (faeces) or stomach contents are limited in that they cannot provide an accurate quantitative reflection of food items consumed by the seals while at sea (Pitcher 1980, Hawes 1983, Bigg & Fawcett 1985, Da Silva & Neilson 1985, Murie & Lavigne 1985). However, a useful index of seal diet can be derived from analysis of otoliths from seats. There are a number of problems in interpreting the results obtained from seats due to digestion of the hard parts used in identifying the prey species. Broken or digested otoliths can give incorrect estimates of the age/length classes of prey consumed . If total digestion or vomiting of some items occurs, they might be totally missed in the analysis and in general, there might be a bias against large items (Pitcher 1980, Hawes 1983, Da Silva & Neilson 1985). Nevertheless, scat analysis can still provide an idea of the range of species taken, relative abundance and some information of the size of prey taken . Here I use these parameters to look at the feeding ecology of the fur seals' at Punta San Juan. Fresh seats were opportunistically collected from beach S3 and areas of beaches N4 and N5 (Table 1) known to be utilized almost exclusively by fur seals and only very ocasionally by sea lions. In these three sites fur seals of all age and sex groups are normally present throughout the year, but as most samples come from beach S3 where mainly adult females and immatures occur and because most adult males fast during the breeding period (Chapter 6), the otoliths in the Table 4.1- Scat samples of S.American fur seal collected at Punta San Juan between 1983 and 1985 ----- -- -- ---- -- -------- ---------- -- ---- -- -·------- -- -------·----- ---------------- -- ------ ------ -- ------------ ----------- Year Month Number Samples with otoliths --------------------------------------------------------------------------------------------------------·--------·----- 1983 Jan/Feb 17 ? March 7 ? April 11 ? June 16 ? July 25 ? August 21 16 October 13 6 December 62 13 1984/5 January 41 20 September 27 16 October 21 20 December 16 14 1985/6 277 ----------------------------------------------------------·---------------------------------------------- ? = Samples analysed in bulk, not individually. 23 seats will largely represent the adult females' and possibly the immatures' diet. METHODS: Only freshly deposited seats were collected and placed in plastic bags. Later, they were transferred into plastic containers where they were soaked in a saturated detergent solution (Magia Blanca) for 1-7 days and finally rinsed with water successively through three nested sieves with mesh diameter of 2.0 , 1.0, and 0 .5 mm (the last only in 1985). When most of the soft organic material was washed through, all fish otoliths (sagittae) and squid beaks were removed and stored dry (1983 and 1 ~84) or in water (1985) for subsequent analysis . Most of the otoliths were identified by me by comparison with reference collections. Otherwise, they were identified by Dr.G.A.Antonelis (National Marine Mammal Lab., NW & Alaska Fisheries Center, Seattle), Dr. Juan Velez (Institute del Mar del Peru, Callao) or Ms.Sandy Hawes (SW Fisheries Center, San Diego) . Some items have not been identified yet. Maximum length and width were measured, for all otoliths except those which had sustained damage, to 0.01 mm using a OLYMPUS SZ-111 stereoscopic dissection microscope fitted with a graticule. Digestion was estimated by eye and ranked 1 to 3 (1 = < 20% digestion, 2 = 20 to~ 40 %, 3 = > 40 %) based on the presence or absence of ridges or grooves known to be present in the undigested otolith . Length and weight of whole fish were estimated only from the least digested (stage = 1 ), whole otoliths of anchoveta using the regression equation for length: Lt= 24.96 + 2 9,1 Olt ( 1) Where Lt = total length of anchoveta in mm and Olt = maximum length (corrected for digestion) of otolith in mm x 0.1 (modified from Chirinos de Vildoso & Chuman 1968) . In these estimates, digestion of the larger otoliths (> 3 mm 24 long) was assumed to be l1:3ss than in the smaller ones. Thus, the large otoliths were assumed to have lost only 10% length from digestion and not 20% as mentioned above . Anchoveta weighting (Wt) was calculated using the equation : Wt= 0.007 (Lt/10)3 (2) derived by Dr. Daniel Pauly (pers. comm) . These lengths and weights of anchovetas were subsequently compared with data from the commercial fisheries to determine the age class to which they belonged. For all the other species of prey identified average lengths and weights were obtained from various sources, but the available information did not allow detailed analysis as above. For purposes of analysis, the samples have been split into 12 groups based on the month when they were obtained . Statistical tests used in the analysis of the data are mentioned in the results and were taken from Siegel (1956) and Sokal & Rohlf (1983). To determine the importance of a prey item , two indices were used: the proportion of the total number of otoliths collected that belonged to a particular prey species (here referred to as "number of otoliths") , and the number of seats in which a given prey item was found ("frequency of occurrence"). Both indices were calculated for each collection period in 1984 and 1985. For the 1983 seats, only number of otoliths could be calculated because the samples were not analysed individually but several at a time. Using the number of otoliths as an index might however, overestimate the importance of very small prey. High numbers of otoliths could result from a few individual seats containing large numbers of otoliths coming from very small fish. In the analysis below, I assume that the most important prey item is the one showing the highest frequency of occurrence. If most seats contained a given item, it is likely that most individuals were consuming the same type of prey. RESULTS: 25 Between January 1983 and December 1985, 277 scat samples were collected, from which 1,938 otoliths and nine squid beaks were recovered. The prey species identified, as well as the numbers of otoliths for each species for each period are listed, in order of decreasing number in Table 4.2. Twenty species of prey were identified but 86% of the otoliths belonged to only four species of fish : Leuroglossus urotramus (40%) , anchoveta (34%), sardines, Sardinops sagax (6%) and species number 21 which is still unidentified (6%) . Other prey species contributed less than four% of the otoliths. Variations in diet composition and diversity: Most individual scat samples contained fewer than 1 O otoliths (1984 : med = 3, n=39, 1985: med = 5, n=50, Fig. 4.1). There were no significant differences between years , but the highest numbers of otoliths per scat were found in 1984 when three samples, containing 458, 153 and 106 otoliths of L.urotramus, accounted for 92% of the otoliths of this species and 37% of all the otoliths recovered. Most seats contained only one or two types of prey (77% in 1984, 94% in 1985, Fig. 4.2), but the proportion of samples containing three or more species of prey was greater in 1984 than in 1985 (G = 6.2, df = 2, p < 0.05) . Diet composition varied widely, but in eight out of twelve collection periods anchoveta appears as the most important item consumed (Table 4.3) . However, the numbers and size ranges of otoliths of this species varied between periods. The June and July 1983 samples only included one and four anchoveta otoliths respectively and thus, have not been contained in the calculations below. When comparing the three years studied, the sizes of anchovetas taken by the seals were found to be significantly different (T-test: 83 vs .84 : t = -10 .68, df = 66.2, p«0.001; 84 vs .85 : t = 18.5, df = 103.7, p < 0.001 ; 83 vs .85 : n.s .). No differences in sizes were found within years . When the frequency distributions of total length of anchoveta for each of these periods were compared with the data for age-length distributions from the fisheries (Chirinos de Vildoso and Chuman Table 4.2 - Prey species and numbers of otoliths present in the scat samples of S.American fur seals at Punta San Juan, 1983-85. (Species appearing as numbers have not been identified yet). ------------------------------------------------------------------------------------------------------------------------------------Prey species I 1983 I 1984 I 1985 I Jan ./Feb . Mar. Apr. Jun. Jui. Aug . Oct . Dec. Jan. Sep. Oct . Dec. Total ---------------------------------------------------------------------------------------·--------------------------------------------Leuroglossus urotramus 0 0 0 0 0 0 3 592 167 12 3 0 777 Engraulis ringens 41 27 13 1 4 0 13 55 67 53 277 103 654 Sardinops sagax 2 4 2 10 19 9 1 21 47 2 3 0 120 21 0 0 0 0 0 0 1 72 47 0 0 0 120 Trachurus symmetricus 0 0 0 19 16 23 5 0 0 0 0 5 68 Anchoa spp. 0 0 0 0 46 0 0 0 1 0 0 0 47 Aphos porosus 0 0 0 0 3 0 0 0 31 0 0 0 34 15-Bathylagidae 0 0 0 0 0 0 12 2 0 8 6 2 30 06-Myctophidae 0 0 0 0 1 1 0 7 1 12 5 1 28 Cynoscion analis 0 0 0 1 17 0 0 0 0 0 0 0 18 Odonthestes regia 4 0 9 0 0 0 1 1 0 0 0 0 15 Merluccius gayi 0 0 0 1 3 1 1 1 0 0 3 0 10 Squid 0 0 0 0 0 0 0 0 0 4 5 0 9 16 0 0 0 0 0 0 0 3 3 0 0 0 6 10-Serranidae 0 1 0 0 2 1 0 0 0 0 0 0 4 14 0 0 0 0 0 0 1 0 0 0 1 0 2 Mugil cepha!us 0 0 0 0 0 0 0 2 0 0 0 0 2 09 0 0 0 0 1 0 0 0 0 0 0 0 1 13 0 0 0 0 0 0 1 0 0 0 0 0 1 24 0 0 0 0 0 0 0 1 0 0 0 0 1 -------------------------------------------- ----------------------------------------------------------------------------------------------------Total Number of Otoliths 47 32 24 32 112 35 39 757 364 91 303 111 1947 -------------- ------------------------------------------------------------------------------------------------------·--------------------- >50 • 1984/5 n = 39 50 D 1985/6 n = 50 a, 0. 40 ,._ E 0 IQ I.. (/) a, ....... ..0 V\ E .c. :, ~ 30 z~ 0 +' 0 20 10 0 0 10 20 30 40 50 60 70 Percentage Distribution 4. 1 Frequency distribution of number of otoliths found in individual scat samples collected in 1984 and 1985. .... Cl, 0 a. I... E Cl, ,v .Q "' E '- ::, 0.. z 0.. "' 5 • 1984/5 n = 39 0 1985/6 n = 50 4 3 2 0 10 20 30 40 50 60 70 Percentage Distribution 4.2 Frequency distribution of number of prey species found in individual scat samples collected in 1984 and 1985. Table 4.3 - Prey species" ranked by frequency of occurrence over 1984 and 1985. For each collection period, the occurrence of each species is indicated and the number 1 shows the most important species consumed during that period"" ------------------------------------------------~--------------------~--------------------------------------------------Prey species I 1983 I 1984 I 1985 I Jan./Feb. Mar. Apr. Jun. Jui. Aug. Oct. Dec. Jan. Sep. Oct. Dec. Total --------------------------------------------------------------------------------------------------------------------------Engraulis ringens 1 1 1 X X 1 1 X 1 1 1 11 Sardinops sagax X X X X X X X X 1 X X 11 Leuroglossus urotramus X X X X X 5 15-Bathylagidae 1 X X X X 5 06-Myctophidae X X X X X X X 7 21 X X X 3 Trachurus symmetricus 1 X 1 X X 5 Squid X X 2 Merluccius gayi X X X X X X 6 16 X X 2 Aphos porosus X X 2 Odonthestes regia X X X X 4 14 X X 2 Cynoscion analis X X 2 Anchoaspp. 1 X 2 10-Serranidae X X X 3 13 X 1 Mugil cephalus X 1 24 X 1 09 X 1 Number of species /period 3 3 3 5 10 5 10 11 8 6 8 4 • Species given as numbers have not been identified yet. •• Rank based on frequency of occurrence, except for 1983 when it is based on number of otoliths because the samples collected that year were not analysed individually . 26 1968), the 1983 and 1985 samples fitted the three year old (or older) anchoveta distribution while the 1984/5 sample fitted the length distribution of one year old anchovetas (Fig . 4.3) . In 1983 and 1985, the length-frequency distribution of anchoveta consumed by the fur seals closely matched the fishery catch composition for the same time periods (Fig. 4.4) . Diet diversity varied with the most important item taken. In the periods when adult anchovetas (three or more years old) were the most important prey (Jan .-Mar. 1983 and Sept.-Oct. 1985, Table 4.3), the diet was less diverse (x = 4.5 ± 0.8 species of prey/period , n = 6) than when young anchovetas (one year old) or other species of prey were the most important prey items (x = 8.2 ± 1.1 species of prey/period, n = 6) . This difference was significant (Mann-Whitney U = 4.5, p < 0.025). Even though diet composition and diversity changed within and between years, the length and weight ranges of the species of fish which comprised the fur seals' diet had features in common (Table 4.4). All the most important prey items were less than 30 cm long, weighed less than 300 g and were either epi- or mesa-pelagic species known to carry out vertical migrations at night (Fitch & Brownell 1971; Gj0saeter & Kawaguchi 1980; Jordan 1971 ; Longhurst 1971 ). The few demersal species recorded also fell within similar length/weight ranges and all are known to leave the bottom after dark ( Leible et al. 1981). Relationships with availability of prey at sea Are variations in the presence of adult anchoveta in the seats indicating changes in the abundance or availability of this resource? No direct measurements of prey availability were collected in this study. However, because seals prey on anchovetas of the same age-length groups, within the same depth ranges and during the same times of day (see Chapter 5) as the local fisheries (Jordan & Chirinos de Vildoso 1965). and because all other species of prey occupy a very similar niche, what we know from the anchoveta fisheries enables 45 40 35 30 25 20 15 10 5 0 Scat Samples: lilll 1 983 n = 35 • 1 984 /5 n = 87 D 1 985 /6 n = 273 -~--0-+-:1()1..-+-lO--+-.-:,J:::::~-.__,flw ·•- 1 Year old anchoveta ·0- 3 Year old anchoveta i: . 70 80 90 1 00 11 0 1 20 1 30 1 40 1 50 1 60 1 70 Total Length (mm) 4.3 Frequency distribution of total length of anchoveta (columns)estimated from otoliths recovered from scat samples in 1983, 1984/5 and 1985/6. These are compared with length distributions of known- aged anchovetas obtained from fisheries data (lines) (from: Chirinos de Vildoso & Chuman 1968). Young anchoveat predominate in 1984/5. C 30 .Q ·i 20 E 8 10 "if. 6 30 V) g_ E 8 "cl- C 0 V) 0 20 10 40 30 o. 20 E 0 U 10 "if. 0 1983 1984-85 1985 7 8 Jan.- July Oct., Dec. 8 Jan. Oct., Dec.a Jan. Se pt., Oct. 8 Dec . Sept . ,Oct. 8 Dec 9 10 11 12 13 14 15 16 17 18 19 Body length (LT,cm) 4.4 Length-frequency distribution of anchoveta (shaded columns) eaten by fur seals in different years, compared with fishery catch composition data (--) for sampling periods as closely matched as possible (unpublished fishery data from IMARPE, courtesy of I.Tsukayama). Table 4.4 - Prey species of the S.American fur seal at Punta San Juan 1983-1985 Prey species • Family Comrron Habitat Length Wegrt name (cm) .. (g) -------------------------- -------------------------------------------~--------------------------------Engraulis ringens Engraulidae Peruvian anchovy Sardinops sagax Clupeoidae Sardine Leuroglossus urotramus Bathylagidae Deepsea smelts 15 Bathylagidae Deepsea smelts 06 Myctophidae Lanternf ish 21 Trachurus symmetricus Carangidae Jack mackerel Squid Merluccius gayi Merluccidae Silver hake 16 Aphos porosus Batrachoididae Pez fraile Odontesthes regia Atherinidae Silverside/smelt 14 Cynoscion analis Sciaenidae Cachema Anchoa spp. Engraulidae Anchovy 10 Serranidae 13 Mugil cephalus Mugilidae Mullet 24 09 Habitat: P = Pelagic, MP = Meso-pelagic, D = Demersal, + = Carries out vertical migrations • Species appearing as numbers have not been identified yet. •• Values are ranges or upper limits (-30 = up to 30) p + 6-19 -30 p + 15-30 150-200 MP + 2-7 <10 MP + 5-10 MP + 10-20 p + 30-35 300-500 p + D +- -30 -250 D +- 10-15 p 15-20 30-50 D 15-20 -400 p +? 6-19 30-50 D -35 -300 27 some tentative conclusions to be drawn. The periods when adult anchovetas were missing from the fur seals' seats were related to periods when they were not available to the fisheries . Adult anchovetas did not occur in the seats during the winter in 1983 (Jun.-Aug .) and during the spring and summer of 1984/5 (Oct .-Jan), when they took only very young anchovetas. Environmental conditions in these two periods were very different but in both cases it is possible to relate them to a low availability of adult anchovetas to the fisheries . During El Nif'lo , in June 1983, the few anchoveta schools that had been present in January / February were either dead or had migrated south (W.Arntz, pers.comm) . Also, during the winter, even in "normal" years, anchovetas become less available as the warm water front moves away from the coast (Chapter 3) . With the information available, it is not possible to separate the El Nit'lo effects from the normal winter dispersion . In 1984/5, no anchovetas were being taken by the fisheries around Punta San Juan (Pedro Llerena, pers .comm.) while the seals were taking young individuals of this species, probably the product of the spawning of the few survivors of El Nit'lo. Anchovetas normally only live 3-4 years (Jordan & Chirinos de Vildoso 1965), and after the very severe El Nit'lo of the previous year, it is likely that the surviving adults died after spawning and hence were absent from the seals' diet . The periods when adult anchoveta did occurr in the diet (Jan.-April 1983 and Sept.-Dec. 1985) can be related to a relatively high availability of anchoveta to the seals at Punta San Juan. In March, early on during the 1983 El Nino , a survey of the Peruvian coast carried out by the Sea Institute (IMARPE) showed that the Q..oly anchoveta patches were found near Punta San Juan, close to shore (IMARPE, unpub.data) and in 1985 the IMARPE surveys reported a recovery of anchoveta along the Peruvian coast (J .Alheit, pers. comm) . DISCUSSION: Studies of the Northern fur seal (Callorhinus ursinus, Kajimura 1984) J 28 and of the California sea lion (Zalophus californianus, Antonelis et al. 1984, Bailey & Ainley 1982) for which concurrent data on food availability exist, have shown that the relative proportion of a specific prey item in the diet (as estimated from stomach and scat samples repectively) reflect availability and abundance of that item at sea. This apparently is also the case for the fur seals at Punta San Juan. Although the information on changes in availability of anchoveta for the Peruvian coast is limited , the data here presented suggest that there is a relationship between the fur seals' diet , as described from faecal material, and prey abundance and availability, and that S.American fur seals at Punta San Juan prey almost exclusively on small pelagic fish, apparently with pref ere nee for adutt anchovetas. In 8/12 periods anchoveta was the most important (most frequently occurring) prey. Unfortunately, little can be said about the biomass of fish consumed because of the low number of otoliths found in most seats and because one scat may represent more than one meal. Only few seats contained more than two types of otoliths, thus apparently fur seals usually take only one or two prey items per meal. Since they prey mainly on small schooling fish, perhaps once they encounter a school of a given species of fish, they feed on that school until they fill their stomachs. The seats might represent the last meal before they came ashore. Diet diversity was apparently related to whether or not adult anchoveta was available at sea. In the periods when adult anchoveta was not the most frequently occurring item in the seats, a wider range of prey was taken and these coincide with the periods of reported anchoveta scarcity. In contrast, when adult anchoveta was available to the fisheries, fur seals took few other prey (84% of all the otoliths recovered from the seats in 1985 were from 2-3 year old anchoveta) . An increase in the diversity of species taken when the preferred prey item was not available has been shown to occur in the California sea lion (Bailey & Ainley 28 and of the California sea lion (Zalophus californianus, Antonelis et al. 1984, Bailey & Ainley 1982) for which concurrent data on food availability exist, have shown that the relative proportion of a specific prey item in the diet (as estimated from stomach and scat samples repectively) reflect availability and abundance of that item at sea. This apparently is also the case for the fur seals at Punta San Juan. Although the information on changes in availability of anchoveta for the Peruvian coast is limited, the data here presented suggest that there is a relationship between the fur seals' diet, as described from faecal material, and prey abundance and availability, and that S.American fur seals at Punta San Juan prey almost exclusively on small pelagic fish , apparently with preference for adult anchovetas. In 8/12 periods anchoveta was the most important (most frequently occurring) prey . Unfortunately, little can be said about the biomass of fish consumed because of the low number of otoliths found in most seats and because one scat may represent more than one meal. Only few seats contained more than two types of otoliths, thus apparently fur seals usually take only one or two prey items per meal. Since they prey mainly on small schooling fish, perhaps once they encounter a school of a given species of fish, they feed on that school until they fill their stomachs. The seats might represent the last meal before they came ashore. Diet diversity was apparently related to whether or not adult anchoveta was available at sea. In the periods when adult anchoveta was not the most frequently occurring item in the seats, a wider range of prey was taken and these coincide with the periods of reported anchoveta scarcity. In contrast, when adult anchoveta was available to the fisheries, fur seals took few other prey (84% of all the otoliths recovered from the seats in 1985 were from 2-3 year old anchoveta) . An increase in the diversity of species taken when the preferred prey item was not available has been shown to occur in the California sea lion (Bailey & Ainley 29 1982) . Thus, despite the dramatic decrease in biomass of anchoveta off the Peruvian coast since 1973, fur seals apparently still show a preference for anchoveta. Changes in the availability or abundance of this resource affect the fur seals . If anchovetas are scarce, they have to take a wider range of prey. Also , during El Nil'lo events, even if adult anchovetas are available in sufficient density, fur seals will obtain less energy from the same number of fish because the condition of anchoveta is poor, resulting in a lower calorific content (Pauly & Tsukayama 1987). 30 CHAPTER 5 - FEMALE FORAGING BEHAVIOUR Fur seals prey mainly on anchovetas or similarly sized epi- or mesopelagic fish species, which undergo daily, seasonal and interannual variations in their patterns of vertical distribution, density and abundance (Chapter 4). How do these changes in availability affect the seals' foraging behaviour? In this chapter I will address this question by looking at the dive patterns of lactating females using instruments which provide a detailed continuous record of the diving activity (timing, depth and duration of dives) of seals at sea (Kooyman et al. 1983) . Changes in the dive pattern of seals, assuming the purpose of dives is to obtain prey, may reflect changes in prey availability (Croxall et al 1985, Gentry et al 1986a). Here the dive patterns of the South American Fur seal at Punta San Juan are compared to the known changes in vertical distribution of anchoveta (Chapter 3). METHODS: Dive recorders Diving behaviour of adult female fur seals was studied using Time-Depth- Recorders (TDRs) . These recorders have been describe.d in detail elsewhere (Kooyman et al. 1983, Gentry & Kooyman 1986b). The TDRs weigh approximately 500 g and are housed in cylindrical aluminium pressure cases 20 cm long and 5 cm in diameter (Fig. 5.1) . Dive depth is recorded onto film by a light emitting diode (LED) fixed to the arm of a pressure sensitive Bourdon tube. This arm moves across the film through an arc which is proportional to the changes in pressure. The film is moved past this arm and spooled by a small electric motor. Time is recorded on the film by a second LED fixed to the chassis of the instrument which produces a small dot every 12 min (Fig. 5.2). The TDRs used in this study were fitted with Bourdon tubes sensitive up 5. 1 Female fur seal carrying a time-depth-recorder (TDR) . • y- SWIMM ING """1' • • • • DIVING • • • • \ ~ TIME DOTS...._. • I!:... .. • -.~ • ~RESTING~ • • • • • 5.2 Dive profile as recorded by a TOR (from Feldkamp 1985). Time dots represent intervals of 12 min. Vertical distance from top of each peak to baseline is proportional to depth of dive . • 31 to 300 psi (about 206 m) and in theory could resolve dives greater than 1 O m in depth and 30 sec in duration. Resolution at shallow depths though, was different for each of the four records recovered . Units with low pressure ranges such as the ones deployed in this study (some can record up to 1000 m, Kooyman 1981), normally tend to produce a "chatter" around the baseline in response to the animal's activity . The width of this chatter directly affects the instruments' capacity to resolve shallow dives and it depends on the degree of "stiffness" at which the LED arm is fixed before deployment. Usually it is fixed to allow a moderate chatter so that variations in the width of the baseline can be used as an indirect measure of the animal's activity . If the arm is allowed to chatter too widely , shallow dives falling within the range covered by the chatter are obscured. This happened in both of the 1985 records . If the arm is fixed so that no chatter is produced, the distinction of activities other than diving is impossible but resolution is greatly improved and even dives shallower than 10 m can be distinguished. The recorder used in 1984 was fixed in this way. In the case of the recorder used in 1983, the arm was fixed to produce a moderate chatter. However, in this record the baseline gradually shifted upwards. The problems of interpretation caused by this shift in the baseline and by the differences in resolution between records will be discussed elsewhere . Here, when comparing between records, one must bear in mind that in both the 1"985 records the proportion of dives to depths shallower than 40 m might be greater than the ones given here. All instruments were calibrated before and after deployment by connecting the units to a portable pressure sta1ion and producing marler Duraton Dives Mean Max. Mean Max. (h) (m) (mn) --------- -------- -- -- -- ------- -------------- -- ---- -- ------ -- -- ------ ---------------------·----------------------- --------------------1 24 83 54 170 1.9 5.7 2 47 190 56 95 1.8 3.8 3 8 58 43 87 1.2 2.4 4 115 665 43 101 1.9 6.3 7 1 24 177 19 52 2.1 4.0 2 15 131 10 46 1.5 3.1 3• 166 817 25 126 2.1 4.7 5 1 33 238 18 105 1.2 5.6 2 44 341 25 116 1.1 8.0 F 1 34 342 26 97 2.0 9.3 2 48 440 31 114 1.9 7.7 ·Film ended before female came ashore . 38 CHAPTER 6 - BREEDING BEHAVIOUR AND ANNUAL CYCLE This chapter describes the general characteristics of the breeding cycle and behaviour of female and male S.American fur seals at Punta San Juan. As in all other fur seals , S.American fur seals are polygynous and sexually dimorphic (Bonner 1981) . Adult territorial males weigh approximately 180-200 kg (King 1983) and adult females weigh between 30-90 kg (see below) (Fig. 6.1) . Since the behaviour of the fur seals in Peru is in general very similar to that of other fur seals (Miller 1975 a & b, Stirling 1971 a&b), here I shall describe briefly the breeding and social behaviour of the fur seals during the breeding season. I concentrate on those aspects of the behaviour of the fur seals at Punta San Juan which might be affected by breeding in a hot environment, and see if living in a subtropical environment has an effect on their breeding system. METHODS: Timing of Births: The number of pups born each day was determined from the daily counts of live and dead pups on beach S3 (Chapter 2) using the following formula Where n = day number, 8 = number of pups born on day n, L = number of live pups counted on day n, and D = number of dead pups counted on day n. It was assumed that the number of fresh pup carcasses that were censused early in the morning reflected the number of pups that died throughout the previous day and night (double counting was avoided as described in Chapter 2) . But some pups, which died early in the morning (after the census) and were scavenged by Band-tailed gulls (Larus be/cheri) and Turkey vultures throughout the day, may have been missed by the counts and a few fresh carcasses may have been washed away by the surf; thus the values given here are likely to 6. 1 Large territorial fur seal bull surrounded by females and a pup, showing marked dimorphism in size between adult males and females. 39 underestimate the number of pups born during the early part of the breeding season . Later in the season , when the number of dead pups exceeded the number of pups born (see Chapter 11), the above formula produced negative values although there were still many pups being born on the beach . To overcome this problem, B was estimated as twice the number of placentas counted in the early morning census (Bn(2P)L from the week when negative values were first obtained . This new estimate was obtained from the regression of the estimated number of pups born each day on the number of placentas (Fig. 6.2) . For the three weeks previous to the one where negative values of Bn were first obtained, these two values were not significantly different (t=-1.407, p=0.1688, df=33). Thus, Bn(2P) was taken to give a similar approximation to the number of pups born each day and was used thereafter. The peak of pupping for each year was taken to be equal to the median pupping date . For the analyses below, the breeding season was arbitrarily divided into three unequal parts: early (before 5% of the pups were born), late (after 95% of the pups were born) and mid-season for the period in between . Observations of births and copulations These were carried out mainly during the 1984 breeding season simultaneously with the mother-offspring time-budget study and only on beach S3 (see Chapter 8). Throughout the day (06:00 to 18:00) the study area was constantly scanned and every birth and copulation recorded . Every time a birth or a female in labour was observed, the following information was taken : time of day, position in the rookery, local substrate, whether the pup was born alive or dead, if it was born head or breech first, time until the afterbirth was expelled and if a yearling was present or not. For every copulation observed, the time, location, substrate, and duration (until the bull dismounted) were noted, as well as whether the male was territorial or sub-adult, whether or not the female struggled throughout the whole copulation, and whether or not the copulation appeared to 100 • 90 ,-.. 80 i::: • co • ....; 70 z • Cl::'. 0 60 • (D • Q • w I- 50 ~ • • ~ I- 40 (/) w • 30 20 10 0 0 5 10 15 20 25 30 35 40 NUMBER OF PL AC ENT AS 6.2 Relationship between daily count of placentas on beach S3 and estimates of number of pups born the same day (B , see p. 38) n Y = 1.71 Sx + 8.706, R2 = 0.544, n = 33, p « 0.001 . 40 be successful (if the female struggled all the time and the bull was never seen to thrust regularly for at least five minutes, he was deemed to have been unsuccessful) . Only those days for which there were 12 hours of observation were included in the analyses below. Male territorial behaviour Every morning, between 28 September and 27 December 1985, the study area on beach S3 was checked for individually recognizable males and their location noted . Every time a new bull arrived on the study beach , he was checked to see if he had any distinctive markings (e .g. scars or bites on the fore flippers) that would enable easy recognition . If he did, he was subsequently included in the check list and followed until he was replaced or departed. Some bulls may have been present from before they were first included in the list (i.e. their markings were not immediately noticed), therefore, the duration of tenures given here are minimum values for each individual. For some of these animals, a size category (small, medium or large) was assigned. These were estimated by eye in comparison with neighbouring males. Male activity-budgets were collected for four complete days during the 1984 breeding season. Observations were taken from 06:00 to 18:00 by two observers taking alternate three hour shifts. Each day, three individually recognizable territorial bulls were simultaneously and continuously observed and the number of copulations, sexual (other than copulations, i.e. sniffing) or aggressive interactions with adult females, and the type and number of interactions with other males were summed for each hour of the day. All I-tests mentioned in this chapter are paired and 2-tailed. RESULTS: The annual cycle Adult female fur seals were present at Punta San Juan all year round 41 although the proportion of time they spent ashore (see Chapter 7) and use of some areas varied with season. Fig. 6.3 shows the typical breeding cycle for an adult female Peruvian fur seal. The breeding season at Punta San Juan takes place between October and December. Once a year, females give birth to a single pup and then remain constantly with it for a period of about 8-12 days, at the end of which they copulate and start a series of cycles in which they alternate between foraging at sea and suckling their young on land . Females normally suckle their pups for 11 -36 months but may give birth every year. The variation in the number of adult females on beach S3 throughout the breeding season in 1984 and 1985, and on beach N5 in 1985 is shown in Fig. 6.4a. Beach S3 is the main fur seal breeding beach at Punta San Juan and is used by females all year round. Beach N5 is a much smaller rookery and is occupied by fur seals for only part of the year. Although in 1984 beach S3 was not censused in October, in both years female numbers peaked around late November. The increase corresponds to the build up of females during their perinatal attendance period, and the decrease to the period after most females had given birth and were spending part of their time feeding at sea. In both years the maximum number of females was observed 1-2 weeks after the peak of pupping (see below), but in 1985 it was one week later than in 1984. Also, in 1985 there was a significantly higher number of females ashore on beach S3 than in 1984 (Is= 6.7, P<<0.001, comparison between weekly means) . In both years there were no fur seals on beach N5 at the beginning of the breeding season but numbers increased rapidly from early November onwards, closely matching the changes in numbers takiAg place on beach S3 (Fig. 6.4a) . In 1984, this beach was occupied later than in 1985. In 1985 the first animals arrived on 1 O October, whereas in 1984 there were no fur seals until 6 November. N5 was censused only a few times in 1984, therefore it is not possible to make a detailed comparison between years . Females breeding on this beach were November 1 YEAR i--1 --------(,--------- Birth Copulation ~ 8d t Perinatal attendance period II II 2d 2d Trip to sea 11'11 II • II // II II II II Visit ashore 6.3 Diagrammatical representation of the breeding cycle of female South American fur seals in Peru, showing periods ashore (solid blocks) and trips to sea (gaps). II Birth Copulation ~. t Perinatal attendance period --- ---.. (/) w ...J -<( 1:: w LL. LL. 0 O:'. w en 1:: :::> z z -<( w 1:: (/) w ...J -<( 1:: ...J -<( O:'. 0 t: O:'. O:'. w I- LL. 0 O:'. w en 1:: :::> z z -<( w 1:: 600 550 500 A 450 400 350 300 250 200 150 100 50 0 14 21 28 4 11 18 25 30 7 14 21 28 Odober November December WEEK ENDING 0 S3 - 1984 D N5 - 1984 • S3 - 1985 • N5 - 1985 35 30 B 25 20 15 10 5 0 14 21 28 4 11 18 25 30 7 14 21 28 Odober November December WEEK ENDING 6.4 Seasonal variation in (A) number of adult females and (B) territorial males on beaches S3 and N5 in 1984 and 1985. Each point is the mean value for the week (7 d) ending on the date shown, except for beach N5 in 1984 when the points represent single counts. Dashed line shows period when highest numbers of adult females were recorded. Arrows show the median pupping date. 42 significantly smaller than females breeding on S3 in 1984 (F=39.63, p« 0.001, Table . 6.1) but there was no significant difference in 1985 and 1986. Adult territorial males were present at Punta San Juan only during the breeding season. Throughout the rest of the year there were always a few sub- adult males holding "territories" around the breeding rookeries, but these were soon displaced by the larger males arriving in October. In 1984, because S3 was not censused until November, the number of territorial males did not show any obvious trend (Fig. 6.4b) . There was a small increase between the first and second week of November and then numbers remained relatively constant until the first week of December, when they dropped. In 1985, there was an increase in the number of males from early October until mid-November, but again little variation until after the first week of December, when males started to leave the area. On beach N5 in 1985, the increase continued until the first week of December, but numbers dropped immediately afterwards, at the same time as on beach S3. In all cases, the number of males did not start decre-asing until one week after female numbers started to drop. Like the females, there were more territorial males on S3 in 1985 than in 1984 (t8=-5 .516, P<<0.0001 , comparison between weekly means) . The adult sex ratio (females/territorial male) ranged between 12:1-21 :1 with a mean of 16 females per territorial male throughout November and December, peaking in late November, when most births took place (see below) . There were no significant differences between the three samples (Fig . 6.5, F2=1 .3). Composition of the breeding rookery On both study beaches, apart from the adult females, their pups and the territorial males, there were yearlings, 2-3 year old immatures and sub-adult males present throughout the breeding season (Fig. 6.6) . On S3 50-60% of the animals ashore were adult females (for all percentages given in this section, Table 6.1- Weights* of adult females on beaches S3 and N5 (kg) Year 1984 1985 1986 S3 Mean±S.E 58 .97 ± 1.32 52 .35 ± 2.50 57.48 ± 1.27 n 30 15 50 N5 Mean:±S.E 46.41 ± 1.51 52 .29 ± 2.52 55.15 ± 2.48 n 28 7 10 • All females were weighed within the first 3 days of their perinatal attendance period. 22 21 r--. w 20 .....I < I: 19 '- (/) w 18 .....I z < < I: w 17 I: w I.J... >- '-' .....I Q 16 ::.::: w I- w < ): ~ 15 X w (/) 14 I- .....I ::, 13 C < 12 11 OS3- 1984 es3 - 1985 .&N5 - 1985 4 11 18 November 25 30 WEEK ENDING 7 14 December 21 6.5 Seasonal variations in adult sex ratios (females/territorial male) ashore on beach S3 in 1984 and 1985 and on beach N5 in 1985. Each point is the mean value for the week (7 d) ending on the date shown. w (.'.) < f- z w u 0:: w Cl. w (.'.) < f- z w u 0:: w Cl. w 0 < f- z w u 0:: w Cl. S3 - 84 100 -.-~~~~~~~~~~~~~~~~--, 90 80 70 60 50 40 30 20 10 ...... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .... . . . ... . . ..... . . .. .. . . .. . . o••••.-.....-.-..-...;. ... ..-.-.-~ 4 11 18 25 November 30 7 14 21 December 28 S3-85 1 0 0 177.~;;;;;;;;:;;;;;;;;,;:;;::;:;:;:;;:;;;;;;;;:;;;;;:;;;;:;;;;:;:;;;;:;;:;;3 90 80 70 60 50 40 30 20 10 0 4 11 18 25 November 30 7 14 21 December 28 NS-85 100 ..,..,...,..,...,...,..,...,...,"7"7'7"7"7'7"7"7'7..,,..,.,.77"777"7'Y77"T7T.>"7'7":....,...,..,...,...,"TT7-rT7-rT7'7'T7'1 90 80 70 60 50 40 30 20 10 0 4 11 18 25 November 30 7 'w'EEK END ING 14 21 December 28 • Terr. males D Sub-ad . males 8 T + S-ad . males IS Ad . females ~ 1-3 y immatures D lndeterminates 6.6 · Variation in the age and sex of the fur seals, excluding newborn pups, ashore on beach S3 in 1984 and 1985 and on beach N5 in 1985 throughout the breeding season. 43 newborn pups were excluded) . This proportion remained relatively constant throughout the season , but increased slightly from December onwards, because some yearlings started leaving the area (see Chapter 9). Only 3-4 % of the fur seals on these two beaches were territorial males; this proportion did not vary throughout the season . Yearlings , immatures and sub-adult males together comprised 30-35% of the animals ashore . There were significantly more yearlings on S3 in 1985 than in 1984 (t=7 .9, df=8, p« 0.001, x34=99±5.2, x35=162 ±8.3), but there were significantly more immatures and sub-adult males in 1984 than in 1985 (SAMs: 1=6.9, df=8, P<<0 .001, x34=90±8.9, x35=24 ± 1.5, Imm.: 1=10 .9, df=8 , p«0.001, x84=36±2.7, x85=12±0.98) . These changes might be due to the fact that a different person censused the area each year. The total number of animals ashore on beach S3 ( except pups) in 1985 was not significantly greater than in 1984 (t=2 .1, p=0.06, df=8; x34=602±33, x35=608±23) and in both years, territorial males, sub-adult males and adult females together, comprise the same percentage (71 %) of the animals ashore on S3 . Similarly, yearlings and older immatures together account for about 24% of all fur seals on beach S3 in 1984 and 1985. Thus, the differences in number of adult females , territorial males and yearlings between the 1984 and 1985 censuses might simply be a result of some individuals having been mis-identified as sub-adult males or older immatures . Timing of births The South American fur seal at Punta San Juan is a highly seasonal breeder. Pupping took place from early September to late December each year, but 90% of the births occurred within a period of approximately 40 days, with a median birth date of 11 and 19 November for 1984 and 1985 respectively . The distribution of births in the 1984 and 1985 breeding seasons ( Fig. 6. 7) shows that the 1985 curve is significantly shifted to the right (t64= 3.24, p=0.0019) . The dates by which 5, 50 and 95% of the pups were born (Table 6.2) were later in z O::'. 0 en (/) Q_ ::::> Q_ u.. 0 O::'. w en r: ::::> z w 2'.: I- 4: _J ::::> r: ::::> u 1800 1600 1400 1200 0/ ./-- • 1984 0 1985 1000 800 600 400 ----------------------------------------------~~ : 200 I I I I o~~~~---,-ll~--,----,-~__J 29 9 19 29 8 18 28 8 October November December DATE 6. 7 Cumulative number of pups born on beach S3 in 1984 and 1985. Dashed lines indicate median pupping data. 18 Table 6.2 - Timing of S.American fur seal births at Punta San Juan (beach S3) 1984 1985 First recorded birth 1 Oct. 5 Sep. 9 First successful pup2 Oct. 22 Oct . 26 5% of pups born by Oct. 23 Oct. 26 Median date (peak) of pupping Nov. 11 Nov. 19 95% of pups born by Nov.29 Dec. 12 Total number of pups 1,381 1,704 produced 90% of all pups produced in: 38 d 41 d 1- Observations started later in 1984. 2- First pup that survived for more than 5 days ( see Chap.11) . 44 1985 than in 1984, even though in both years 90% of all births took place over a similar timespan (38d in 1984, 41d in 1985) . Also , in 1985, there were about 300 more pups produced on beach S3 than in 1984. In the two breeding seasons, all pups born early in the season were either stillborn or died within 1-2 days. Healthy pups (those which survived more than 5 d) appeared only after 22 October. In 1984, even though the exact date of pupping for most females was not known, it could still be estimated because all were captured and tagged within the first three days of their perinatal attendance period (determined from their attendance pattern, see Chapter 7) . All females breeding early in 1984 bred again early in 1985, and all 1984 late breeders, bred late in 1985 as well (Fig. 6.8). However 11 out of the 15 females which were observed pupping in both years.pupped in 1985 5-28 days after their estimated birth date for 1984 (x = 11 ± 2.27d) . Only one female gave birth on the same day as in the previous year, and three pupped earlier (6-12 d) . Thus, the difference in the median date of pupping between 1984 and 1985 appears to be caused by a delay in giving birth for most females on beach S3 in 1985. Breeding behavfour of females Birth-s: Before giving birth, females made long foraging trips (see Chapter 7), returning to the rookery 0-2 days before pupping. They usually came back to the site where they gave birth the year before. 21 of 23 females which were tagged on beach S3 in 1984 and observed giving birth in 1985, did so within 30 m of the place where they pupped the previous year. The two females which changed sites had lost their pups in 1984. They were both successful in 1985 and pupped in the new site in 1986. Also 11 out of 14 females tagged on beach N5 in 1984, pupped again on the same beach in 1985. No exchange of females between N5 and S3 was observed . The only beach changes observed took place outside the breeding season, and these involved only short distances: l{) (X) (]\ Q w CL CL :J CL w I- - 0 z I.. a, ..a 28 18 8 28 18 8 29 $ 19 0 0 • • • . ······· ·· ··········· ········· . .. ········· • • • , , •' • ,, •' • ,, • e, ·· • ,•' ,,,,•' • .... • ·················· ,,, ,• ' ,,,, ·· ··· ···•·· 9 ··········· 9 19 October 29 8 18 November DATE TAGGED 1984 28 8 December 6.8 Relationship between date when tagged in 1984 (= pupping date - 3d) and pupping date in 1985 for individually known females on beach S3. Dashed line indicates 1 :1 relationship (females pupped same day as previous year). Note that most females pupped later in 1985 than in 1984. 18 45 females from S3 moved to S4 or S5 and females from N5 moved to the other side of the same beach) . Pupping appeared to follow a bimodal pattern within the daylight hours (07 :00-18:00, Fig . 6.9a) with peaks early in the morning and again late in the afternoon. However, the observed births account for only 40% (range: 27-66%) of the total number of births estimated for the day (Bnl - If 60% of the births take place between 18:00 and 07 :00, then it is possible that they occur at a relatively constant rate from late in the afternoon until mid-morning, with a drop between 10:00 and 16:00. Until night observations are carried out, it will not be possible to determine which is the real pattern. Females preferred to give birth on the intertidal rocks and the wet front part of the beach . They did not start using the back, dry areas of the beach for pupping until half way through the season, when female density was high. Of 126 births, 95% took place on the intertidal rocks, in tidepools, or on the wet gravel along the front part of the beach; the rest occurred on the dry, sandy slope in the back. Of 21 births where delivery was observed, 60% were breech and 40% head first deliveries. Females did not assist the pup during birth . A few pups were born still completely wrapped in the foetal membranes and in none of these cases did the female help the pup out. Most of these died shortly afterwards. If the umbilical cord did not break at birth, they generally turned around or pulled the pup about until the cord broke. If the afterbirth was not expelled with the pup (40% of the cases), it came out between 5 and 200 min. afterwards (x=48±7.1 min , n=44) . Occasionally some females were seen eating their own placentas, but in general they tended to ignore it and allowed the seagulls and vultures to eat them. A few females were seen carrying around their placenta for several days, defending it from scavengers and other seals. Initial mother-pup interactions: ~ 0 :r: ....... (/) :r: I- ~ CD 7 A 6 5 4 3 2 0 07 08 09 10 11 12 13 14 15 16 \ 7 TIME OF DAY 5,--------------------------------, 4 B 3 0 07 08 09 10 11 12 TIME OF DAY 13 14 15 16 17 6 . 9 Variation in mean (± S.D .} of (A} number of births per hour and (8) number o copulations per hour with time of day. Each comlun is the mean of 7 h of observation . 46 The bond between the female and her pup was established shortly after birth. They repeatedly called back and forth to each other and the female spent time smelling the pup during the first hours after delivery. Pups spontaneously searched for their mothers' teats when a few hours old . Although initially they poked with their nose against any available part of their mother (females did not help their pups find the teats), they soon found them and as they grew older, spent increasing amounts of time sucking each day (see Chapter 8) . The perinatal attendance period: Immediately after giving birth females became less tolerant of other seals approaching them and fiercely defended an area surrounding them and their pups. For the first 3-4 days after pupping, mother and pup kept in close contact. At this time, females which had to move around carried their pups in their mouth. During the perinatal attendance period females face two conflicting requirements : 1) to provide their newborn pups with an area safe from other aggressive females and from the surf, and 2) to thermoregulate, which involved leaving the beach and the new pup. At night and while the sun was low, and when the tide tended to be higher, females tried to keep themselves and their pups dry and away from the heavy surf by s1aying in the higher parts of the beach. As soon as air temperatures and radiation started to rise, they had to get wet. Females breeding early in the season, apparently chose to give birth in the gravel next to the intertidal where the distances they had to move during the day were minimal. Later on in the season, as the number of females pupping increased, some were forced to give birth further back on the beach . To get to the water, females breeding at peak densities, had to move across a beach crowded with aggressive females while carrying their pups in the mouth. This caused an increase in pup mortality which is discussed in Chapter 9. Later in the perinatal period, as pups became more mobile, they started 47 wandering away from their mothers . Pups did not aggregate in "creches", they usually remained alone, perhaps playing in a tidepool, never going too far (>30 m) from their suckling site . Even from this early stage, pups were seen to find their way back to where they left their mothers . Towards the end of the perinatal attendance period, if mother and pup were separated either because the pup moved away or the female moved down to the intertidal rocks , it was usually the mother which started calling for her pup in the late afternoon when most females moved up to the back of the beach again (see below) . If the pup called its mother before then, usually there was no response from the female . Pups readily responded to any calling female and females used smell as a cue for further recognition . Females approached by a pup other than their own, would aggressively reject it as soon as they were close enough to smell it. Copulation: The perinatal period usually ended shortly after the female copulated . Of all copulations recorded (n=197), 77% took place in tidepools or shallow water in the intertidal area. Of the remainder, only one took place on the dry, back areas of the beach and the rest either on the rocky parts of the intertidal or on the wet gravel in front of the beach. Copulations could be observed at any time of the day, but they appeared to occur more frequently around noon and early in the afternoon (Fig. 6.9b) . No night observations were made. Non-oestrous females usually rejected any male approach aggressively . In contrast, females in oestrous were commonly seen swimming in a tidepool in front of a territorial male, turning around and presenting their perineum close to the bull's nose. The males usually responded quickly to these signals and mounted the female in the same tidepool. Copulations lasted from 5-45 min. (x=14±1 .7min) and were always ended shortly after the female started biting the male and struggling to move away. There was no obvious sign of females choosing any particular male. However, in the few cases where sub-adult males were observed trying to 48 copulate, the female always struggled from the first contact and usually managed to escape. Only four individually marked females were observed to copulate . They apparently only copulated once, but it is not possible to prove that they did not copulate again later, somewhere out of sight from the observers . Lactation: After mating, females stayed with their pups for 0-2 days and then left them alone while they went to sea to forage, coming ashore to suckle between foraging trips. The females' attendance patterns are described in detail in Chapter 7 and the mother-offspring interactions during the mothers' visits ashore are examined in Chapter 8. Spatial distribution of females: Diurnal movements of females within the rookery were synchronous and predictable . Every day , after 09 :00 when the cliff behind the beach (S3) no longer provided shade , females moved down to the intertidal area to cool off and remained there throughout the hottest period of the day, not returning until around 16:00, when air temperatures dropped and/or the tide came in (Fig. 6.10) . Between 09 :00 and 16:00 (these times varied with cloud cover; on cloudy days females moved down later), the dry, back areas remained devoid of fur seals and almost all females were in the intertidal rocks and tidepools . These thermoregulatory movements were perhaps the most important factor determining the distribution of territorial males and had an effect on pup mortality ( Chapter 9) . Movements of individual females were not random ; they tended to shift between a few preferred rocks or resting sites , associating with several territorial males during the day. Thus, there was not a stable harem structure. However, because of the females' fidelity to specific places on the beach, it is very likely that they associated with the same set of males each day. 6.1 o Daily thermoregulatory movements of fur seals on beach S3. Open water and intertidal area on the left side of pictures, which were taken at three hour intervals from 06:00 (top) to 18:00 (bottom) (GMT-5). Male territorial behaviou r Territory location and size: 49 Territorial bulls showed a stable and homogeneous distribution . All territories were located either along the gravel front, or in the intertidal area (Fig. 6.11) ; there were none at the dry back part of the beach . Within the limited territorial area (approximately 700 m2). there were between 12-31 tenured bulls throughout the breeding season. Because the absolute number of bulls within the same area increased from early October until early November, the size of the territory defended by each bull changed in proportion to the total number of males present. Each bull kept a distance of about 2-4 m between himself and his nearest neighbour (Fig. 6.12) so that each territory covered approximately 20-30 m2. However, males in the gravel area were able to extend their territories towards the back because there were only sub-adult males there which were not able to prevent them from moving up. Thus, these territories could become as large as 100-150 m2. These males only defended intensively the landward side of their t-erritories after 16:00, when most females had moved back up the beach. Duration of tenure: Territorial males showed two distinct 'strategies'. Some had a long, single period of tenure and others arrived early in the season, acquired a territory for a short time and then left and returned later in the season to the same territory for a second tenure of similar duration (Fig. 6.13) . Single tenure males remained on territory between 2-46 days (x=23.22±2.33d , n=37). early arriving males staying longer than those arriving late in the season (F=7 .898 , p=0 .008 , Fig . 6.14) . Double tenure males had a mean first stay of 12 .3±2.6d (range : 3-22d , n=8), a second stay of 12.9±2.6d (range : 3-22d) and a mean absence between the two of 41 .6±8.2d (range : 7-71d) . There was no significant difference between the duration of the first and second tenure for these bulls (t=-0.33, p=0 .75) and the total amount of time spent on territory by single and double tenured males , .... R , ... ., ...... - G , ... B . : ... 6.11 Area of beach S3 covered in the study of territorial males. Open water at the top and cliff at the bottom. HWM = high water mark. Arrow indicates position of observation blind. R = intertidal area, G = gravel area, B = dry back area. 6.12 Spacing of territorail males on beach S3 . Black circles indicate position of bulls . ' 1.-- - - - - - 29 9 19 2 9 0 CT . MEDIAN PUPPING DAY ..., - - - -~ - T 8 1 8 NOV. 28 I - - - - - - - 8 1 8 2 8 DEC, 6 .13 Duration of tenure of individually known bulls (each represented by a horizontal line)on beach S3 in 1985. ,.._ "'O ...__, w Q:'. :::, z w I- LL 0 z Q I- <( Q:'. :::, C 60 • 50 • • • • 40 • • • • • • 30 • • 20 • • • 10 • • • • • • • • • 0 19 29 9 19 29 8 18 28 8 18 September October November December DATE OF ARRIVAL 6.14 Variation in duration of tenure with arrival date of territorial males on beach S3 in 1985. Only males with single tenure included. y = -0.25x + 35 .575 , R2 = 0.184, n = 36, p = 0.008 50 (xn + T2= 25 .13±4.95d, range : 8-44d) was not significantly different (F=0.12 , p=0.73) . Six of the eight double tenure males were seen to be displaced at the end of the first tenure and seven returned to the same territory. Of the 26 males which were given a size category, none of the five considered to be small were double tenure. There was no significant relationship between size and duration of tenure. Site fidelity: Three males with very distinct marks on their fore flippers were identified as occupying the same territory on beach S3 in 1984 and 1985. It is very likely that many more males which were territorial in 1984 were back in 1985, but with the available information, ij is not possible to tell. Site tenacity: Within the 1985 breeding season site tenacity of individual bulls varied widely . Apart from the two 'strategies' described above, there were two shorter term variations in male behaviour. Bulls which held territories either on the gravel or tidepool areas, remained there constantly throughout the day and night for the whole duration of their tenure. These males appeared thin and emaciated towards the end of the breeding season . Smaller males which were seen defending areas at the edge of the main territories, just behind the gravel area or in the shallow water next to the intertidal rocks, tended to abandon their 'territories' during the times of the day when there were no female-s around . Males at the back, which had no access to water, either left the beach through specttic paths across the territories, giving a loud submissive call (Trillmich & Majluf 1981), or walked along the back of the beach leaving the area through a section of the waterline not defended by territorial bulls . They usually left around 10:00, once all females were down in the intertidal area, and returned the same way they lef1, early in the afternoon (around 15 :00) before the females came back . Observations of two individuals indicated that the same small males returned to the same place on the beach every day . These males never appeared to be 51 emaciated, probably because they were able to forag e during the day, while they were at sea. Males holding are as in the shallow water immediately in front of the intertidal rocks, did so only when the tide was low and most females were down by the water (usually between 09 :00-16:00) . When the tide rose or the females went back up the beach, they swam away. While away, these males might have also spent some time foraging , as they always looked in very good condition . It was not possible to check whether the same males returned each day nor whether they were the same as the ones holding areas in the back. These aquatic males often tried, unsuccessfully, to get copulations by jumping onto females arriving or leaving the beach. The intensity with which bulls defended their territories also seemed to vary through the breeding season. When disturbed by researchers either very early or very late in the season, even the largest bulls left their territories without much resistance . From mid-November to mid- December, when most females where giving birth, males would attack fiercely , never leaving their territories . Peripheral males were more easily scared anytime in the season. Male-male interactions: Territory boundaries were maintained by means of ritualized displays and fights similar to those of other fur seals (Mccann 1980, Miller 1975a, Stirling 1971) . Sub-adult males trying to enter territories were quickly chased off by the resident bull. When a large male trying to take over the territory presented a challenge, resident and intruder repeatedly displayed to each other (Fig. 6.15), a fight occurring only if the intruder did not retreat. From the limited data , it appears as though interactions with other males by territorial bulls (chases , displays and fights combined) took place mainly early in the morning and around noon (Fig . 6.16) . Yearling or 2-3 year old immature males were often observed lying right next to the territorial bull without eliciting an aggressive response. The presence 6.15 Boundary display between two territorial males on beach S3 . 12 0:: ::) 0 I 10 ........ (/) z: Q 8 f-- u <( 0:: 6 w f-- ~ w 4 ...J <( 1:: I w ...J 2 <( 1:: 0 6 7 8 9 10 11 12 13 14 15 16 17 TIME OF DAY 6.16 Variations in mean (±S.D.) frequency of male-male interactions with time of day. Each column represents mean of 12 h of observation . 52 of sub-adult males was tolerated although mainly in the back of the beach, where no territories were defended (Fig . 6.17) . These smaller males in the back also had to move to the water around noon, and apparently some managed to stay in the intertidal area without being expelled by the tenured bulls . A few ( 1-5) also managed to stay in the front area, but then it was mainly at the very back of it, around the high water mark. The behaviour of males appeared to indicate that no copulations occured away from the water. Males did not defend territories in the dry back parts of the beach . Males in the gravel area did tend to shift their landward boundaries upwards when females were coming in, but the boundary extensions were only limited to the area below the high water mark (Fig. 6.11 ). Also, sub-adult males were never seen attempting to mount a female while in the back of the beach. The only violent attempts by small males to mount, albeit always unsuccessful , were observed to be carried out by the aquatic males in the shallow waters in front of the intertidal. Interactions with females: Because of their thermoregulatory movements , the number of females on each territory varied widely through the day. Fig. 6.18 shows the variation in the mean number of females per territorial male with time of day for the front gravel area and for the intertidal area. Only around noon did males in the intertidal area have available significantly more females than males in the gravel area (1=4 .96, df=8, p=0.001 ), but on average, the latter had significantly more females available throughout the day (1=5.8, df=8, p«0.0004, paired comparison between means for whole day) . However, these females were by no means restricted in their movements and they walked freely across territory boundaries without any attempt by males to stop them. In 144 hours of observation, males were never seen trying to herd a female. Most male-female interact ions (other than copulations) took place when females were either moving up or down the beach, (f) 25 w ...J <( I: I- 20 ...J ::> C, <( I 15 m ::> (f) LL. 0 10 a:: w m I: ::> 5 z z <( w I: 0 ~ lntertida 1 Iii Front gr ave 1 • Back 06 09 12 15 18 TIME OF DAY 6.17 Variation in mean (±S.E.) number of sub-adult males with time of day in the different areas of beach S3. Each column represents mean of 9 h of observation. 22 w 20 _J < I: a::: 18 a::: w I- 16 ........ CJ) w _J 14 < I: w LL. 12 LL. 0 a::: w 10 (l) I: ::> z 8 z < w I: 6 I: 4 e FRONT 0 INTERTIDAL 6 8 10 12 14 16 18 TIME OF DAY 6.18 Variation in adult sex ratio (females / territorial male) with time of day according to location in the breeding beach. Each point represents mean of 9 censuses. 53 around mid-morning (09 :00) and again late in the afternoon (16 :00, Fig . 6.19) . These involved mainly inspections (nuzzling of the perineum) of any passing female, or open-mouth-threats (Miller 1975a) when a female was found lying in any of the males' favoured resting spots . If two females having an aggressive interaction were noticed by a male, he would normally rush and inspect both . These, and any other approaches by males, were generally responded to aggressively by females . In contrast, apart from when competing for resting spots, males were not aggressive to females. As there was no attempt by males to retain females in their territories, females were never wounded by males, as is common in the Northern fur seal (M .Goebel , pers. comm.). DISCUSSION : Differences between years (1984 vs. 1985) During the 1985 breeding season the median pupping date was later than in 1984. This change in the timing of births apparently was due to most females giving birth later in 1985. For this general delay to occur, the most likely time during pregnancy when most females could have been similarly affectec:l was at implantation, which in fur seals occurs about three months after copulation (King 1983). A correlation between timing of implantation and body fat reserves has already been shown for grey seals (Halichoerus grypus, Boyd 1984) . The delay could thus have been caused by females being in poorer condition at the time of implantation in 1985 than in 1984. Of the four females which were weighed during their perinatal attendance period in 1984 and reweighed at a similar time in 1985, three had lost weight (3-15.5 kg) and the mean weight of females on S3 in 1985 was significantly lower than in 1984. However, a similar weight loss was not observed for the females on N5 and they peaked in numbers at the same time as in S3 (and thus probably had a similar time distribution of births) . Higher female densities could also delay the timing of pupping as has 35 et:'. 30 ::, A 0 I '-(/) 25 w _J < 20 I: w LL. 0 15 f- (/) z Q 10 (/) (/) w et:'. 5 (.'.) (.'.) < 0 6 7 8 9 10 11 12 13 14 15 16 17 TIME OF DAY 60 55 et:'. 50 B ::, 0 45 I '- (/) 40 w I 35 u < 30 0 et:'. CL 25 CL < 4-0 _J < 15 :::> X w 10 (/) 5 0 6 7 8 9 10 11 12 13 14 15 16 17 TIME OF DAY 6.19 Variation in mean (±S.D.) frequency of (A) male-female aggressive interactions and (B) sexual approaches with time of day. Each column represents mean of 12 h of observation . 54 been shown to take place in red deer ( Cervus elaphus, Clutton-Brock et al. 1987). However, with the available information, it is impossible to determine what caused the variation in pupping dates in 1985. In 1985 there were more pups produced and apparently, the age-sex composition on beach S3 was different : there were significantly more adult females , territorial males and yearlings , than in 1984. The differences in composition could be related to differences in identification of certain age/sex classes . It is not possible with the available information to determine what caused the differences in pup production . Differences between breeding beaches (S3 vs. N5) Beach NS is a much more accessible beach and human disturbances are more common than on beach S3. Also, because it is very open and has a wide, gentle guano slope in the back, N5 tends to be used by sea lions at different times of the year. Sea lion sub-adult males frequently disturb the corner of the beach used by fur seals when they come to steal pups (sometimes as frequently as 3-4 times a day) . These and the human disturbances might be responsible for the desertion of this beach during part of the year. The only beach at Punta San Juan that is used by fur seals all year round is S3, which has a single point of access (for humans) and is too rocky for sea lions which prefer sandy beaches. Females which breed in areas like NS, outside the breeding season (after May) move to rock outcrops and islets out of reach from humans. They have to move to more protected areas to pup, but they remain there only while their pups are small(< 6 months). More than 60% of the fur seal population at Punta San Juan breeds on S3. Thus, it seems that S3 is the breeding area preferred by fur seals . Facto;s affecting reproductive success of females and males The main factor constraining the behaviour onshore of breeding male and female fur seals at Punta San Juan seems to be the need for access to water for thermoregulation . Adaptations to avoid hypothermia while at sea, adversely affect thermoregulation while on land (Chapter 12) . At high air temperatures, \' 55 avoiding the sun and getting wet are among the most effective behavioural mechanisms to maintain thermal balance in otariids (Whittow et al. 1971 , Umberger et al. 1986) . At Punta San Juan, on the two study beaches, availability of covered areas is very reduced as there are only a few small caves and large boulders providing shade at the back of each beach. Therefore, the main means of thermoregulation for most animals is becoming wet . For females, if they give birth in the back dry areas of the beach, they are forced to make daily movements to and from the water's edge . During the perinatal period these movements increase the probability of mortality of newborn pups, specially around the peak of pupping. The effects of these movements on pup mortality are discussed in detail in Chapter 9. For males, the most important factors determining the location of territories appeared to be 1) availability of females and 2) access to water for thermoregulation . If females were not receptive while in the dry back areas of the beach, then, by controlling access to water, tenured males gained exclusive access to oestrous females . Thus, there were no territories defended on the back of the beach, without access to water; they were held only along the front gravel and in the intertidal. Males holding territories in the gravel area had access to females under all thermal conditions , tidepools for thermoregulation, and the possibility of extending their ranges towards the back at times when females moved up. Males in the intertidal area had access to high numbers of females only around noon and had no flexibility in the size of their territories . The relative mating success of males in these two types of territories still needs to be investigated. Male-female interactions Territorial males did not attempt to control the. movements of females moving through their territories, but let them go by with a quick inspection. This lack of control of males over females' movements appears to be related to the 56 high densities and to the diurnal movements of females . By herding, males are able to control only one female at a time. In a situation like the one on beach S3 , where movements of females tend to be synchronous (all are affected by increased solar radiation at the same time), any attempts by males to restrict female movements are bound to be ineffective. Instead, males locate their territories along the areas which are most likely to be used by females throughout the day. Female choice No obvious signs of female choice were observed, but by copulating only in the wet areas, females could guarantee being mated by the dominant males which were able to hold territories . Rejections of sub-adult males could also be a form of mate choice. Aggressive interactions between females and sub-adult males, could be a way of attracting the attention of the resident male so that the sub-adult male is chased away. Female aggression to intruder males as a mechanism of mate choice has been suggested for grey seals (Boness et al. 1982) . The same study also suggests that in otariid species, where oestrous is very short (1 -2 days). female solicitation could serve to ensure that oestrous does not pass without insemination. In the case of S3, where female density is very high, such a mechanism would be of particular value, since males cannot keep track of all females passing through their territories. By soliciting, females would not only ensure insemination, but would also guarantee being mated by a prime male. Territorial tenure The mean duration of tenure for males at Punta San Juan was relatively short, when compared to other species of otariids (Table 6.3) . Male body size might play a role in determining intra-specific differences in duration of tenure (Trillmich 1984) , but it seems unrelated to inter-species differences. Thus , California sea lion and Galapagos fur seal males had the same mean duration of Table 6.3 - Duration of tenure and body weight of territorial males in several species of otariids. Species Latitude Callorhinus ursinus 1 57N A.gaze/la 2 52S A.pusillus 3 30S A.australis (PSJ} 15S A.galapagoensis 4 0 A.forsteri 5 35S A.townsendi 6 28N Eumetopias jubatus 7 59N Zalophus californianus 8 30N Weight of males (kg} 182-272* 140* 200-350* 136-159* 70** 200· 1 ooo· 300* !:X.Jration ofterure (d} 47 34a 42 23 27b 37 30 43 27 Double terure refX)rted X X ---------------------------------------------- ------------- ------------------------------------------------------------------- Location of study & source of data: 1 - Pribiloff Is, Bartholomew & Hoel, 1953 2 - South Georgia Is., Mccann, 1980 3 - S. Africa, Rand, 1967 4 - Fernandina, Galapagos Is., Trillmich, 1984 5 - South Neptune Is., N.Zealand, Stirling, 1971 6 - Guadalupe Is., Baja California, Pierson, 1978 7 - Montague Is ., Alaska, Sandergren, 1970 8 - San Nicholas Is., California, Odell , 1972 • Weights obtained from King, 1983 •• Gentry & Kooyman, 1986 a - for males seen to copulate, less for non-copulating males b - mean first tenure. 57 tenure, even though the latter are 6-7 times smaller; S.American fur seal fur seal males in Peru, which are three times larger than Galapagos males, had even shorter tenures. The spread of the breeding season might have an effect on the territorial behaviour of male fur seals (Fig. 6.20 and Table 6.3) . The Galapagos fur seal breeding season is very long (90% of the births occur in 10 weeks; Trillmich 1987) and no male was able to hold a territory throughout the entire breeding season. Thus, males being able to return for a second period ashore were able to stay on land for a longer total period of time and presumably were able to impregnate more females. For males with two tenures, the first tenure was longer than the second (Median = 27 and 15d respectively, Trillmich 1984). The difference between tenures was interpreted as a preference by the Galapagos males to mate with early pupping females, which were more likely to produce a surviving pup (op. cit.) . California sea lion males in San Nicholas Is., California, also had a very long breeding season (about 12 weeks long) and two of 16 bulls were observed returning for a second period ashore (Odell 1972). Similarly, no males could stay ashore for the entire length of the breeding season, but the ones holding territories late in the season were more likely to be present when most females came into oestrous and presumably, the ones able to regain a territory were able to increase the number of matings they obtained. The spread of the breeding season for the Northern fur seal in San Miguel was similar to that of the S.American fur seal at Punta San Juan (4-7 weeks; York 1987), but in this case, the number of times a male left the rookery during a breeding season decreased with age . Younger males spent several periods ashore during the breeding season while older bulls had a long, single period of tenure (Delong 1982) . Although this study does not compare the total amount of time on territory obtained each year by any given male, it shows that young males maintained territories in less preferred areas. The Punta San Juan males were able to cover only about half of the Latitude j May June July Aug. Sep . Oct . Nov. Dec . Jan. C.ursinus (Pribiloffs) 1 60N (S .Miguel) 1 30N A .townsendii 2 28N A .ga lapagoensis 3 0 A .austr a lis (Peru)4 15S (Uruguay)5 30S A.philippii 6 30S A .pusillus 7 30S A .p .doriferus 8 30S A.forsterii 9 45S A.tropicalis 10 <52S A .gaze lla 11 >52S 1 - York. 1987 2 - Pierson, 1 978 3 - Trillmich, 1 984 4 - This study 5 - Vaz- Ferreyra, 1987 6 - Torres, 1987 3 - 4 4-7 6 10 5 - 6 4-5 7 - David & Rand, 1986 8 - Warneke & Shaughnessy, 1986 9 - MaHlin. 1 981 1 0 - Kerley , 1 987 11 - McCann, 1980 6.20 Timing of breeding seasons and pupping synchrony in different species of fur seals. Numbers above bars indicate period (in weeks) over which 90% of pups are born . 58 spread of the breeding season and only large males had two tenures. The fact that when males had two tenures , the total amount of time ashore was not significantly increased, and both tenures were of similar durations, suggests that some males might be choosing to breed early and late in the season, rather than holding a longer single period of tenure covering the peak of pupping date . Females giving birth around the median pupping date had a lower probability of producing a surviving pup than those pupping late in the season (see Chapter 9) . Because female fur seals are very synchronous in their pupping (Fig . 6.7) , the offspring of males holding a territory at this time are very likely to be born at the same time the next season, when their probability of mortality is high. Thus, by being able to hold a territory at both ends of the season, double tenure males would perhaps be siring pups with a higher probability of survival. 59 CHAPTER 7 - ATTENDANCE BEHAVIOUR OF FEMALES In this chapter I examine the behaviour of individually known females to assess the possible effects of the female's reproductive status, body weight, and changes in food availability on their attendance at the breeding beaches. The proportion of time that female fur seals spend on land versus at sea while rearing their young determines the proportion of time they dedicate to acquiring food and the rate at which females deliver milk to their offspring. Several different factors are likely to affect the balance between foraging time at sea and time spent ashore. Fur seal females in Peru suckle their young throughout the whole year, sometimes for up to three years continuously, and they may have more than one young sucking simultaneously (Chapters.8 & 11) . Thus, their energetic requirements are likely to vary according to the number and age of their dependent offspring . How females meet these requirements is directly related to the amount of food available in the environment and the efficiency with which they can forage . Increased energy demands on the female by pregnancy or lactation, or a decrease in food availability, might be reflected in longer foraging trips at sea. While on land, females spend their time suckling their young. Therefore, the amount of time females are able to spend ashore might be a function of the size of fat reserves on arrival and on the rate at which the offspring deplete these reserves. METHODS: Attendance behaviour of females was studied in 1983, 1984/5 and 1985/6. Throughout these years, female-offspring pairs or trios in which both the females and their young were marked or where only the pups were marked (see Table 7.1 for sample sizes) were checked at least once and up to four times each day to assess the duration of the females' presences at the colony. Table 7.1 - Dates and mumber of female-days and individuals observed between 1983-1986. (FO ~ both female and offspring marked, F = only female marked, Y = only yearling marked, P = only pup marked). Year Dates 1983 10 Jan./12 Feb. 07 Mar./03 Apr. 12 Apr./09 May. 08 Jun./21 Aug. 1984/5 05 Oct./17 Dec. 13 Jan./17 Feb. 1985/6 09 Sep./28 Dec. 05 Jan./16 Jan. Number of Female-Days 164 118 218 141 641 3,144 828 3,972 2,961 468 3,429 Number of lrdivid.Jals FO F Y P 4 4 1 1 3 1 1 4 1 6 10 7 9 16 5 2 17 10 5 15 15 9 2 12 -------------------------------------------------------------------------------------·------ ------·- TOTAL: 8,042 60 These durations were determined according to the following assumptions. 1.) When seen ashore during the last check one afternoon and the first check the next morning females were assumed to have stayed overnight. 2.) When they were not seen for more than 24 hours (or three consecutive checks), they were assumed to be absent. Arrivals or departures were assumed to have taken place in the middle of the interval between the last check when they were present (or absent) and the following observation (i.e . if a female was last seen in the 0600 check and the next check was at noon, she was assumed to have left at 1 OOO). Nocturnal arrivals or departures were deemed to have taken place at midnight. For the analysis of the attendance records, females have been divided into four categories : - Lactating (L) : females still suckling a yearling at the beginning of the breeding season. - Non-Lactating (NL): - Pupping ( P): females who either did not produce, lost or had already weaned their previous year's pup so that they were never observed suckling a yearling . females observed to give birth sometime during the breeding season. - Non-Pupping (NP): females not observed to give birth throughout the breeding period. The number of individuals and/or the number of trips included in the results below changed according to the subjects being examined . Where because individua l variation wa s · small, groups of females or different years were being compared,vall trips or visits of all the fem ales in the data subsets involved were combined. When changes in the attendance patt-erns of a given female (or group of females) were being investigated, mean durations of trips or visits were computed separately for 61 each of the individuals involved. Only samples of four or more trips or vists were considered in the analyses below. Statistical tests used were the Mann- Whitney U and Wilcoxon's Signed Rank (T or Z) . For sample sizes in Mann-Whitney U tests, refer to Table 7.2. In order to determine how the estimates of durations of female presences or absences from the rookery were affected by whether both the mother and young were marked or only the pup was marked, the results obtained from mother-pup pairs and from records of pups were compared (only post-partum NUP females were considered). Only the estimates of visits ashore turned out to be significantly different. In 1984, the duration of female visits calculated from observations of mother-pup pairs were shorter than the ones obtained from just the pups (U=3775 .5, p=0.0494; M-P: x = 1.5±0.045 d, P: x = 1.8±0.086 d) but in 1985 they were longer (U=7620, p=0.0334; M- P: x =1.6±.068 d, P: x =1.5±0.05 d). The results for January/February 1983 were obtained in collaboration with Dr. Fritz Trillmich (Max Planck lnstitut fur Verhaltensphysiologie, Seewiesen) and have been already published in Trillmich et .al. (1986) . Data for the remainder of 1983 were published in Majluf (1985). Data for July/August 1985 was kindly provided by Pamela Saunders and Ken Darcy (Dept. of Zoology, Cambridge). RESULTS: Non-lactating, pupping females Pupping had a marked effect on the duration of the females' foraging trips. Trips to sea (TS) and visits ashore (VA) were more irregular in duration and significantly longer before giving birth (TS: T = 0, p = 0.003, n=10; VA: T =7, p=0.01, n=11, Fig. 7.1a), while absences at sea lasted about one third as long after pupping (Table 7.2). Mean duration of the perinatal attendance period was 9.2 ± 0.61 d (n=39) and the first trip afterwards was significantly A II B II C II D II II II Ill II II II II - A : Non-lactating, pupping female B : Lactating, pupping female II II Ill II C : Lactating, non-pupping female D : Non-lactating, non-pupping female Birth Ill Ill II PAP Birth II ---- II II PAP Ill II II Ill II II -- 2d II = Periods ashore D = Periods at sea PAP = Perinatal attendance period 7 .1 Diagrammatical representation of the attendance patterns of female South American fur seals in Peru. II Ill II II II II II II II II II Table 7.2 - Mean duration of trips to sea and visits ashore (in days) of S.American fur seal females during the breeding season (Oct.- Dec.), divided by reproductive status and by year. (L c Lactating, NL = Non-lactating, P = Pupping , NP = Non-pupping) [Values= Means (S.0./n)) TRIPS TO SEA: Status Year Pre-partum First Trip Post-Partum NL/P 84 2.9 (1.71/17) 4.0 (2.09/123) 85 9.6 (5 .26/28) 2.8 ( 1 .82/23) 3 .3 (1 .64/210) UP 84 6.6 (1 .69/6) 3.9 (2.32/2) 3.1 (1 .31/12) 85 5.8 (2 .30/73) 2.4 (1 .05/6) 2.8 (0.87/26) L/NP 84 5.4 (2.88/32) 85 4 .9 (1 .93/89) NL/NP 85 6.8 (3.53/16) VISITS ASHORE: Status Year Pre-partum Perinatal Post-Partum NL/P 84 7.6 (3 .02/16) 1.9 (0.86/140) 85 2.2 (0.69/27) 10.3 (4.00/23) 1.6 (0.85/235) LIP 84 1.6 (0.46/6) 3.5 (0.73/2) 1.1 (0 .34/14) 85 1.5 (0.56/71) 8.2 (3.40/7) 1.4 (0.48/32 L/NP 84 1.4 (0.51/37) 85 1.4 (0 .57/92) NUNP 85 2.3 (1.46/17) 62 shorter than the subsequent trips (Z=-2.89, p=0 .0019, n=39) . Otherwise, there were no significant trends with time in the durations of trips or visits during the breeding season. If the pup died, females continued attending the breeding beach, although irregularly. In this case, their presences and absences appeared to be longer but this difference was not significant. Three females abandoned the area shortly after their pups died . Post-partum trips and visits were significantly longer in 1984 than in 1985 (TS: U=9746, p=0.0002, VA : U=12637, p=0 .0001) although the perinatal period in 1984 was significantly shorter (U=98.5, p=0.0146). Lactating, pupping females The attendance pattern of lactating, pupping females was similar to that of non-lactating, pupping females (Fig. 7.1 b). Here again, the attendance pattern changed after the females gave birth. Trips and visits before pupping were longer than after pupping (TS: T =0, p=0 .014, n=6: VA: T =2, p=0.021, n= 7) but significantly shorter than those of pre-partum non-lactating, pupping females (TS: U=525.5, p=0 .000; VA: U=445.5, p=0 .000). There was no difference between the post-partum trips of these two groups of females, but there was a significant difference between their post-partum visits to the rookery (U=27602.5, p=0.0002) . Mean duration of the perinatal attendance period for lactating, pupping females was 7.2±1.2 d and in this case, the first trip after birth was not significantly shorter than subsequent trips. Again, there was no trend with time in the durations of the females' presences and absences during the breeding season. After their mothers had given birth, the yearlings continued nursing simultaneously with their younger siblings for variable periods of time. Of the ten females included in this group (two in 1984/5 and eight in 1985/6), two weaned their yearlings immediately after pupping, four lost their pups at birth 63 and continued suckling their yearlings, and four nursed both young until January 1986, although only one was still doing so by March (see Chapter 10) . If nursing the yearling and pup simultaneously, or yearling only if the pup died, her attendance pattern did not vary significantly from the overall pattern of post-partum, pupping females, except that the stays on land of the latter were significantly shorter (U = 5085.5, p=0.0005, x = 1 .24±0.07 d, n = 30) . Apart from the post-partum visits in 1984, which were significantly shorter than in 1985 ((U=145, p=0.0421), there were no major differences between the 1984 and 1985 lactating, pupping females' patterns. Lactating, non-pupping females Lactating, non-pupping females kept a constant and regular attendance pattern throughout the breeding season (Fig . 7 .1 c) resembling the pattern of lactating, pupping females before giving birth (Fig. 7.2a) . When comparing these two patterns, the pupping females stayed at sea for slightly though significantly longer periods (U=4281, p=0.0055) but their attendance periods were of similar durations. In general, trips and visits of lactating, non- pupping females were significantly longer and shorter respectively than those of post-partum, pupping females regardless of whether they were also suckling a yearling or not (TS: U=9603, p=0.0000, VA: U=20387, p=0.0268) . There were no significant differences between 1984 and 1985. Non-lactating, non-pupping females Although these females had no dependent young, they still kept visiting the rookery throughout the year, alternating short and long trips with visits of varying durations (Fig. 7.1 d) . Their attendance behaviour was less regular than that of any of the previous categories although it was similar to that of pre-partum non-lactating, pupping females (Fig. 7.2b) . The mean duration of trips of the pupping females was significantly longer (U=168 .5, p=0.0196) but there was no difference between the duration of their visits to the rookery. 30 25 w 20 D < I- 15 z w u Ck'. w 10 n.. 5 0 30 25 w 20 D < I- z 15 w u Ck'. w n.. 10 5 0 B D NL/P - PRE • NL/P - POST ·•- NL/NP 2 3 4 5 6 7 8 9 10 11 12 13 14 15 DURATION OF TRIPS TO SEA (d) !\ D LIP - PRE • LIP - POST ·•- L/NP •__J:"1•-+-•-+-• 2 3 4 5 6 7 8 9 10 11 12 13 14 15 DURATION OF TRIPS TO SEA (d) 7 .2 Frequency distribution of duration of trips to sea of (A) lactating females and (B) non-lactating females. Comparisons are made between lactating (L) and non-lactating (NL) females which pupped (P) pre- and post-partum and females which did not pup (NP). In b o th cases NP females show similar patterns to those of pre-partum P females 64 Seasonal changes There was a strong seasonal variation in the duration of the females' absences from the rookery (Fig. 7.3a) . During the summer (Jan-Feb) in 1985, once the breeding season was over, females which were suckling a pup returned to the rookery after significantly shorter periods at sea (U= 2824, p=0.0147) than previously. In the winter (Jui-Aug) their absences were again longer (U=614, p=0.000) ; by September/October, the trips to sea of these lactating females before pupping were on average 1.5 days longer (U=1140, p=0.0002) than earlier in the winter. After giving birth , trips became short again (U=525.5, p=0.000) . The duration of the females' visits also varied throughout the year. Females paid the longest visits to the rookery immediately after giving birth in 1984 and 1985 and also early in the winter (Jui-Aug, Fig. 7.3b) . El Nifio Attendance behaviour under El Niflo conditions was studied between January and August 1983. In January/February, when El Niflo conditions were most severe, females suckling young pups (1 -3 months old) showed very erratic behaviour. They made short and long trips without any pattern, but made significantly more trips which were longer than four days compared with females of similar reproductive status observed at the same time of the year in 1985 when conditions at sea appeared to be normal (X2=6.8, p LL 0 z Q I- < et'. :::, Q z < w I: 0 167 60 47 71 39 BREED. S 84 JAN /FEB 85 JUL / AUG 85 PRE-P ARTUM POST-PARTUM POST-PARTUM BREED. S 85 (OCT /DEC) (OCT /DEC) TIME OF THE YEAR 7.3 Seasonal variation in the duration of (A) trips to sea and (8) visits ashore of female fur seals at Punta San Juan during 1984 and 1985. 30 25 LLJ (!) 20 -<( I- z 15 LLJ u ~ LLJ 0.. 10 5 0 • 1983 n = 21 D 1984 n = 92 mill 1985 n = 84 2 3 4 5 6 7 8 9 10 11 12 DURATION OF TRIPS TO SEA (d) 7.4 Frequency distribution of duration of trips to sea of female fur seals at Punta San Juan in 1983 (El Nino), 1984 and 1985. Only trips of NUP females observed in the summer (Jan.-Feb.) are included. 65 year during El Nit'lo (Fig . 7.5, see Chapter 3 for a detailed description of the variations in SSTs through El Nit'lo) . The highest SSTs and longest durations and greatest variation in the times the females spent at sea were recorded in January/February By March, SSTs at Punta San Juan decreased and trips became shorter (x = 3.67±0.31 d, n=13) and more regular. In April, as temperatures went up again, trips became longer (x = 7.93±0.66 d, n=13) although this time they were more regular in duration than the ones in January/February . Very few short trips were recorded at this time . By July/August, although SSTs decreased with the receding El Nit'lo, females were stili making quite long trips (x = 6.19±0.63 d, n=8) albeit shorter than the ones recorded in April/June. Body weight Female body weight appeared to affect the amount of time a female spent ashore. Heavier females stayed on land for longer periods of time (Fig. 7.6, r=0 .73, p<0.01, n=14) . Duration of trips to sea appeared not to be related to body weight. Only the 1984 post-partum NUP females were included in this analysis because not enough weight information for 1985 or for the other groups of females was available. In 1985, females were lighter, and perhaps had smaller fat reserves, than in 1984 (see Chapter 3) and this might be related to the shorter duration of their post-partum visits in 1985 mentioned above . DISCUSSION : The amount of time a female spent foraging at sea and the duration of her vists ashore varied according to the age and number of her dependent offspring, with her reproductive status and body size and with changes in the availability of food at sea. It has been suggested that the duration of foraging trips and visits ashore are ultimately regulated by the state of their adipose 0 ,,..__ C! u (/) 0 ~ .._, f- C (/) ,v Q 24 22 f .. '•, 20 18 16 14 10 9 8 7 6 5 4 3 2 Jan/F~b n = 21 Mar/ Apr 13 ,' ,, ,, ·· ...... Apr/Jun 16 TIME OF YEAR ' ••, .... .... ···· .. .... '••,,, Jun/ Aug 8 7. 5 Variation in duration of trips to sea with time of year and with sea surface temperature (SST) during El Nino in 1983. 2.2 2 ,....._ "Cl '-/ w Ck'. 0 1.8 I (/) -<: (/) !::: 1.6 ~ > L1.. 0 1 .4 z 0 f- -<: Ck'. 1.2 :::) Cl z: -<: w L . 8 40 • 45 50 55 60 65 70 WEIGHT (kg) 7. 6 Variation in duration of visits ashore with female weight. y = 0.027x - 0.048, R2 = 0.538, n =13, p = 0.0028 75 80 66 tissue depot or fat -lean ratio (Gentry et al.1986a). If so , the duration of visits would be determined by the amount of fat reserves the female has available for lactation and the rate at which energy is removed from the fat stores, which depends on demand of her young . No direct measurements of energy transfer between mother and young were obtained in this study. However, if we assume that 1) yearlings have a higher energy demand than pups because of their larger body size (Costa & Gentry 1986, Chapter 8), 2) that the fat reserves of females with two young sucking simultaneously will be exhausted faster than by having just one young sucking at a time, and 3) that fat reserves scale with body size of the mother (Calder 1984, Lindstedt & Boyce 1985), the results obtained appear to support the idea of a nutritional threshold limiting the duration of females' stay on land which is reached faster with increasing demand of the offspring . Females without dependent young (NUNP and NUP pre -partum, see methods for abbreviations) were able to stay ashore for the longest periods, followed by females suckling only a young pup (NUP post-partum) . Females suckling a yearling (UNP) or females with a pup and yearling sucking simultaneously (UP post-partum) made the shortest visits. This sequence corresponds to an increasing energy demand to the female. Also, heavier females, which have absolutely more fat to turn into energy, were able to stay ashore for longer. Further evidence for this threshold is provided by studies of the attendance behaviour of some other species of fur seals . In the Galapagos fur seal, the only other species of fur seal where yearlings are regularly seen sucking, shorter visits of mothers of yearlings were correlated with longer times sucking and a greater energy demand (Trillmich 1986a) . In the Northern fur seal, when under experimental conditions pups were not allowed to join their mothers until some hours after her arrival ashore (equivalent to a decreased demand) , or pup demand was increased by removing milk from their stomachs , the duration of the female's visits were longer and shorter 67 respectively. In the same species, non-suckling females remained ashore for longer than lactating females (Gentry & Holt 1986). Apart from the perinatal attendance period, the duration of visits varied within a narrow range (0 .5-3 .Sd) and individual females were very consistent in the times they spent on land . The greatest variability was shown by females suckling no young at all. The attendance pattern of these females was the least predictable of all in that neither their trips nor visits were of regular duration . A similar variability in the patterns of non-suckling females was observed in the Northern fur seal (Gentry & Holt 1986) . Time at sea appeared to be the most flexible component of a female's attendance behaviour. Foraging trip duration varied within a wide range (1 - 15d) and these variations seemed to relate to changes in offspring demand and in food availability at sea. Normally, the amount of energy stored should be able to cover the female's metabolic requirements while fasting ashore and should also meet her offspring's demand so that it can maintain itself and grow during its mother's absence. An increased offspring demand under constant foraging efficiency or increased effort related to lower food availability, should therefore be reflected in longer times at sea. The fact that females suckling yearlings made longer trips to sea than females suckling pups supports the idea that trip duration is regulated by offspring demand. If this is so, females suckling both a pup and a yearling should make even longer trips than females suckling only one young at a time. Why then did these females make short trips resembling those of females just suckling a pup instead of longer trips as would be expected? The funct-ion of these short trips in terms of newborn pup survival appears to be clear. As opposed to pups whose only source of maintenance is their mothers' milk, yearlings can partially feed by themselves and because of their larger bo·dy size, can survive without food for longer. The amount of energy a pup 68 receives during a visit is proportional to the amount of fat in its mother's milk. It has been shown that there is a significant correlation between fat content and the duration of the mothers' absences from their pups in several species of otariid seals (Trillmich & Lechner 1986) which suggests that the fat content in the milk of these seals is adapted to sustain a pup's normal activities while fasting. If so, and if the average trip length of females suckling just pups corresponds to the optimal amount of time a newborn or very young pup can maintain itself and grow normally during a maternal absence, then females with a pup and yearling sucking simultaneously might have to reduce the duration oi their trips to sea in order to ensure the survival of the younger offspring even when having to sustain a greater milk demand. Late pregnancy also seems to increase the duration of a female's foraging trips. Immediately prior to giving birth, females made significantly longer trips than females of similar reproductive status but which did not produce a pup. This may have resulted either because of the increased energy demands of pregnancy or because females were trying to accumulate extra reserves to keep themselves throughout the fasting perinatal period. A seasonal increase in trip duration has been described for the Northern fur seal (Gentry & Holt 1986) and for the Antarctic fur seal (Doidge et al. 1986), both studies suggesting a relationship with increasing offspring demand and/or decreasing food availability in the environment. No similar trend was evident for the Peruvian fur seal within a similar period (0-4 months) but later, in the winter, trips were significantly longer and became even longer just before pupping when pups of the previous season were 10-12 months old. This long term increase in trip duration in the fur seals in Peru also could be related to either an increase in offspring demand, a need to accumulate reserves for the coming bree.ding season, or a winter decrease in food availability (see Chapter 3). With the available information it is not possible to 69 separate these effects. The lack of an increase with time in trip duration in the first four months after pupping in the Peruvian fur seal could be explained by looking at the development of this species and conditions at sea in the summer. By the time pups are four months of age (Jan .-Mar.), they might have reached 75% of the weight of a yearling (Chapter 10) , but by then, not only is their milk consumption still low (Chapter 8) , but it is also the time of the year when food availability for females is highest (Chapter 3) . Together, this may explain the short, regular trips throughout the first four months of life of a pup. At this age, pups of the two temperate species mentioned above are just about to be weaned, weigh almost as much as a yearling S.American fur seal, and their milk demand is at its maximum (Doidge et al. 1986, Gentry & Holt 1986, Gentry et al. 1986). Thus, these faster growth rates appear reflected in a rapidly changing attendance pattern. Finally, in the previous chapter, we saw that under El Nif'lo conditions females had to make a greater effort to get to their prey by having to make a larger number of deep dives. However, increased effort does not necessarily imply increased foraging success. Females under these abnormal conditions might still not have been able to obtain enough food even at the greater depths they were reaching. If so, in order to futtill their youngs' and their own energy demands adequately, they would have to spend longer periods at sea foraging. Under normal conditions during the summer, females suckling pups made very short and regular trips (1-6d) . At the same time of the year, under El Nino conditions , females made long trips (>4d long) more often and trip duration varied widely (1-1 Od). This variability suggests a dispersed but patchy distribution of prey. As shown in Chapter 5, some females were feeding close to shore, perhaps on the anchoveta patches trapped by the warm waters offshore. They probably could return to shore sooner. Females that did not encounter one of these patches maybe the ones making longer foraging trips , 70 either because they had to travel longer to get to a patch, or because prey density was so low that they had to spend more time foraging trying to cover their requirements . The variations in trips length observed after March 1983, when El Nit'lo conditions were receding, might be due to prey changes typical of the season (Chapter 3), or to the females trying to regain condition after months without food . CHAPTER 71 8 - LACTATION Young fur seals start feeding independently, while still sucking, at the age of 6-8 months. Until then, they are totally dependent on their mothers' milk. However, their mothers are not continuously available to their young throughout lactation, but only during their short visits in between foraging trips (Chapter 7). Therefore, the amount of milk a female delivers to her offspring during a visit ashore should be enough to allow maintenace and growth of the young during the periods she is at sea foraging . Do females vary the amount of resources they allocate to lactation according to the increasing energy demands of their growing offspring? In this chapter I will try to answer this question by looking at the behaviour of female fur seals and their offspring using the duration of suckling bouts and proportion of time spent sucking by young of different ages as indices of the females' energetic effort. Here I shall also look at the interactions of female fur seals and their young to determine whether there is any conflict over the amount of milk a female is willing to give and the amount of milk her offspring tries to obtain (parent-offspring conflict, Trivers 1974) and also, whether conflict of interests between females and their young varies according to the female's status (i.e. if she is suckling a pup, a yearling, or a pup and yearling simultaneously). To ass-ess these questions, the frequency and direction of aggressive interactions betwe-en mother-offspring pairs (or trios} and the proportion of suckling attempts by her offspring a femaie rejects , will be used as indices of parent-offspring conflict; assuming that conflict of interests is signalled by a decrease in successful! sucking attempts and increased female aggression. METHODS: The behaviour of females suckling young of different ages was studied using one-zero sampling (Altmann 197 4, Martin & 9ateson 1986), recording 72 every two minutes whether the young was sucking or not. Time intervals were monitored with an acoustic digital stopwatch (Timex - quartz) . Observations were carried out on beach S3, from blind A (see Chapter 2) , from 0800-1100 and 1400-1700, from 2-24 November 1985 (see Table 8.1). Observations at other times were not possible because of the demands of other work (Chapter 2). Each session , 6-16 mother/young pairs (or trios) of known history were observed simultaneously, and their behaviour recorded on check-sheets . Only animals which were within an area where they could be easily scanned from the observation blind were chosen . If the time sample for any given pair was shorter than 60 min (some females departed before the end of the session), it was discarded. The proportion of time spent suckling was calculated by dividing the number of checks in which the young was seen sucking by the total number of checks obtained for the pair in the session. A suckling bout and its termination was defined as follows: a) a minimum of two consecutive checks when suckling was observed constituted a bout, b) its termination was defined as at least five consecutive checks (= 1 O min) when suckling was D.Q1 observed. The number of checks (including sucking and non- sucking) between these pauses was multiplied by two to obtain the duration of the bout (in min.). If a bout had already started when the session began or had not ended when the session finished, it was discarded from the analysls below. For comparisons between age groups, the information for all individuals in the age group was combined. When looking for trends within age groups each individual was considered separately, as long as sample sizes allowed it i.e. it was observed in six or more sessions. Rejections of sucking attempts and aggressive interactions between females and their young were studied separately. During days whe.n the above time-budgets were not being collected, within the same time limits, each of the Table 8.1 - Dates and number of female-young pairs (or trios) observed during the suckling behaviour studies. Date T1111e• Femalew~h: Total Pup Yearlirg Pup+Yearl. ----------------------------------------------------------------------------------------------------------------------------------2/11 am pm 3/11 am pm 4/11 am 5/11 am pm 6/11 am pm 7/11 am 12/11 pm 13/11 am pm 14/11 am pm 15/11 am pm 17 /11 am 19/11 am 21 /11 am pm 22/11 am 24/11 am Total: • am: 08:00-11 :00 pm: 14:00-17:00 5 5 6 6 7 5 6 5 5 4 10 8 12 10 10 8 8 8 13 8 7 8 8 172 2 7 2 7 2 8 2 2 10 2 2 11 1 3 9 1 3 10 1 2 8 1 6 1 1 6 1 11 2 10 1 13 2 12 5 15 2 10 4 1 13 1 2 11 3 16 3 11 3 10 2 10 2 10 38 24 234 73 three hours was divided into 20 minute intervals and in each of these, one of the three categories of animal association (female/pup pair, female/yearling pair, or female/pup/yearling trio) was observed continuously , and all aggressive interactions and sucking attempts recorded. For each aggressive interaction, the identity of the receiver (the pup or the yearling) was recorded . For sucking attempts, the success or failure were assessed by whether or not the attempt resulted in sucking for at least five minutes . These observations involved mainly unknown individuals. A total of 626 female-hours of observation were obtained , including 57 different female/young pairs (or trios, Table 8.2). The number of male and female young sampled in each age group was not significantly different from the 1 :1 ratio (M/F: Pups : 20/13, G=1.1, n.s.; Yearlings : 9/12, G=0.19, n.s.), and similarly the difference in sex composition between the two age groups was not significant (G=0.99, n.s.). The number of sessions per individual for female/pup pairs was greater (median=5) than for the female/yearling pairs (median=1, Table 8.2), because most female/pup pairs were observed throughout the perinatal attendance period (PAP) . RESULTS: Female/pup pairs The median percentage of time spent sucking by pups was 6.4% (n=172, Fig. 8.1a) . No suckling occurred in 35% of all pup-sessions . Median duration of suckling bouts for pups between 0-34 days old was 16 min (range: 4- 86 min, n=100, Fig. 8.1b) . However, both , the percentage time spent sucking and duration of suckling bouts significantly increased with age (Fig . 8.2) . Because the samples used to calculate these regressions included non- independent points, pups which were sampled six or more times were tested to see if they showed the same tre-nd. Five out of the six pups tested had positive slopes for both parameters, but due to the small sample sizes (6-13 samples per pup), only two were significantly correlated with age. Table 8.2 - Suckling scans 1985. Number of female-hours Number of individual pairs (or trios) observed Median number of sessions per individual (Range) Mean observation time (min) per individual per session (S.E.) Ferrales wlh: Pups (P) Year1irgs (Y) P+Y Total 464 33 5 (1 -13) 161 .8 (2.7) 96 21 1 (1-5) 152.3 (5.7) 66 3 9 (3-12) 164.3 (8.0) 626 57 >-~ 30 w => 0 Cl::'. LJ.. w 20 (:) < 1- z w ~ 10 w a.. 0 ~ 40 z w => 0 ~ 30 LJ.. w (:) < 20 1- z w u Cl::'. ~ 10 0 Med = 6.4% n I 112 10 20 30 40 50 Med. = 55.4% n = 39 ! 60 70 PERCENT AGE TIME SPENT SUCKING Med. = 36 min n = 67 J. 10 20 30 40 50 60 70 80 90 100 110 120 DURATION OF SUCKLING BOUTS (min) A 80 90 100 • PUPS 0 YEARLINGS 180 190 8.1 Frequency distribution of (A) percentage time sucking and (B) duration of suckling bouts for pup and yearling fur seals. Based on 464 female-h of observation for pups and 96 female-h of observation for yearlings . ('.) ~ ~ u :::) (/) 1- z l.LJ CL (/) l.LJ ~ 1- l.LJ ('.) < 1- z l.LJ u O::'. l.LJ CL 100 80 60 0 • • 5 10 A • • • 15 20 25 30 35 AGE (d) c: 90+--~-....._-~--'--~----'---~----'---~---'--~----J'--~~----1-! 80 (/) I- :) 0 CD ('.) ~ _, ~ u :::) (/) u.. 0 z Q I- < O::'. :::) 0 70 60 50 40 30 • • • • • 201--~~_I__..LJ~~L-........ ----....--- • I • I: • 10 • B • • • • I • • • 0+--~-----~-~-~-~-~-~---~--~-~----1-0 5 10 15 20 25 30 35 AGE (d) 8.2 Variation with age of (A) percentage time spent sucking and (B) duration of suckling bouts for pups 0-35d old. Based on 464 female-h of observation. A) y = 6.104 - 0.104x + 0.037x2, R2 = 0.168, n = 152, p = 0.0001 B) y = 18.43 - 0.551x + 0.043x2, R2 = 0.151, n = 88, p = 0.0001 74 When the mother started foraging at sea (after the perinatal attendance period), on the first day ashore of her return from a trip , the pup spent a significantly higher proportion of time sucking and made longer suckling bouts than on the second day after her return (Table 8.3) . The time spent sucking and suckling bouts during the second day of attendance were not significantly different from those observed for pups during the perinatal period. To determine whether there was any difference between the sucking behaviour of male and female pups, the sample was divided according to whether the pups were still in their perinatal period or not, and for those observed after the perinatal, whether their mother was on her first or second day of attendance. No significant differences between the sexes were found for any of these groups. Some of the pups in the sample died within a month of birth (see Chapter 9) . Their behaviour was compared to that of the pups which survived at least until the following March (1986) dividing the sample as above . No significant differences were found . Female/yearling pairs Yearlings spent a significantly higher proportion of time sucking (med.=55.42%, n=39) and had longer suckling bouts (med .=36 min, n=-67) than pups (Fig. 8.1, %S: M-W Z=-9.11, p«0.00001, SB: M-W Z=-5 .67, p«0.00001). As with the pups, the time their mother had been ashore after a foraging trip also affected the sucking behaviour oi yearlings . They had significantly longer bouts (Medians : d1 : 50 min, n=33 ; d2 : 30 min, n=34 ; M-W Z=-2.74, p=0 .0031) and spent more time sucking (Medians: d1 : 63 .34%, n=21 ; d2 : 50 .84%, n=18; M-W Z=-1.75, p=0.0401) during the first day of attendance than on the second day. There were no significant differences between the sucking behaviour of male and female yearlings. Female/pup/yea·rl+ng trios Only three female/pup/yearling trios were observed : one (female 827) 75 could only be followed for two days because the pup died; two (females 107 and 716, Table 8.4) were observed throughout November. In most of the sessions where these two females were sampled (75% and 78% respectively), the yearling was away or was just lying next to the female and did no sucking throughout the session . Of the two pups however, female 716's pup did not suck in only one session (11 %) while 107's pup did not suck in four (33%) of the sessions . Female 716's pup spent significantly more time sucking (med=26.4%, n=9, M-W Z= -2.541, p=0.0057) than female 107's pup. Neither of these two pups showed an increase in percentage time sucking or duration of bouts with age. In contrast to the yearlings without siblings, the sucking behaviour of the two yearlings was not significantly different from that of the other pups nor from that of their own younger siblings (%S: fem. lli: med.=12.8%, n=4, fem.1.QI: med.=49 .5%, n=4; SB: fem.fil: rned .=30 min, n=4, fem.1.Q.l: med,= 43 min, n=4). In the case of female 827, the yearling was seen sucking in all three sessions they were observed, although making only short suckling bouts (med.= 11 min, n=4). The pup died after the second day, even though it was seen to suck in all sessions, spending a median of 15% of the observed time sucking (the yearling spent 17%). During the sessions where yearling and pup were seen sucking (for all three trios, n=9), only once were both seen sucking simultaneously. To assess the proportion of time females 716 and 107 spent with their pups, yearlings or wtth both together, during November and December 1985 when they were checked daily as part of the attendance studies (see Chapter 7), the number of checks when they were seen ashore with either or both their offspring was calculated. Both females spent most of their time ashore (on beach S3) with the pup only (49 and 47% respectively, Table 8.4). However, when sucking the yearling only , they sometimes moved to the nearby beach, S4, Table 8.4 - Suckling and attendance behaviour of two female/pup/ yearling trios. Female: Table 8.3 - Percentage time spent sucking and duration of sucking bouts of pups during the perinatal attendance period (PAP) and during subsequent visits ashore (divided into first and second day attendance). PERCENTAGE TIME SUCKING: PAP Afterwards Mann-Whitney U test: PAP vs. afterwards d1 vs. d2 PAP vs.d2 day1 day2 DURATION OF SUCKLING BOUTS: PAP Afterwards Mann-Whitney U test: PAP vs. afterwards d1 vs . d2 PAP vs .d2 day1 day2 n 119 21 32 z -3.252 -2 .253 -1 .554 n 62 15 23 z -3 .114 -3.38 -0 .949 median 2.2% 19.1% 9 .2% p 0.0005 0.0122 n.s. median 14 min 30 min 16 min --------------- p 0.0009 0.0005 n.s. Table 8.4 - Suckling and attendance behaviour of two female/pup/ yearling trios . Number of observation sessions Total observation time (min) SUCKLING BEHAVIOUR: Number of sessions when pup was not observed sucking Number of sessions where yearling was not present or did no sucking Median percentage time sucking (n) p y Median duration of suckling bouts (min (n)) p y ATIENDANCE BEHAVIOUR: Number of attendance checks female seen ashore Percentage checks: • with pup - with yearling · with pup + yearling together - alone STATUS BY MARCH: p y 107 12 2,014 4 9 Female: 4.0 (12) 12.8 (4) 27 (8) 30 (4) 57 49 19 10 8 dead still sucking 716 9 1,478 7 26.4 (9) 49 .5 (3) 19 (10) 43 (4) 59 47 .5 13 .5 23 still sucking weaned 76 where it was more difficult to find them . Thus, the amount of time spent with the yearling only may have been underestimated . Female 716 was seen with both young signif icantly more often than female 107 (G=S.64, p<0.02) and was never seen ashore by herself as female 107 often did (14% of the time). Female 107 lost her pup in January and continued suckling her yearling at least until March. Female 716 was still suckling her pup by March, but the yearling was never seen after February. Female/offspring interactions Females rarely rejected sucking attempts by their own young if they were suckling either a pup only or a yearling only (P : 4%, n=24, Y: 8%, n=178), but if they were suckling a pup and a yearling together, they tended to reject both young with the same frequency (38%) and significantly more often than single- offspring females (p>0.01, Fig. 8.3) . There was generally little aggression between females and their young. In most observation periods, no aggressive interactions were observed (70%, n=184) . Aggressive interactions were usually limited to open-mouth threats (Miller 1975a) and only resulted in termination of suckling when the female was about to depart on a foraging trip. Aggressions from the female to the young were generally related to changes in suckling sites. When the fem ale was about to move to the intertidal area to thermoregulate (see Chapter 6), and the young wanted to continue sucking, she would aggressively reject it. Yearlings objected more than pups to their mothers' movements and thus suffered more aggression (Fig .8.4) . They emitted a very loud, high pitched call ('Tantrum' call) every time the female wanted to move . As soon as the female found a new resting spot however, all aggression stopped and suckiing was resumed shortly afteTwards. Aggressive interactions occurred more frequently between the female and her yearling if she was suckling a pup as well (Fig. 8.4) . These again, were limited to open-mouth threats, and rarely resulted in termination of suckling. In the latter case , females would threaten their yearlings without any apparent 0.4 39 107 (/) z n = 65 .33 obs . hours Q I- 0.3 u w J w Cl:'. lL 0 0.2 z Q I- Cl:'. 0 0... 0 .1 0 178 Cl:'. 0... 24 0.0 I 30 I 31 15 15 PUP YEARLING (PUP) (YE ARL) PUP + YEARLING FEM ALE W 1TH : 8.3 Proportion of sucking attempts rejected by females already suckling pups, yearlings or a pup and yearling simultaneously. Figures above columns = total number of sucking attempts observed. Figures at bottom of columns = number of individuals observed. Table 8.3 - Percentage time spent sucking and duration of sucking bouts of pups during the perinatal attendance period (PAP) and during subsequent visits ashore (divided into first and second day attendance). PERCENTAGE TIME SUCKING: PAP Afterwards Mann-Whitney U test : PAP vs. afterwards d1 VS . d2 PAP vs.d2 day1 day2 DURATION OF SUCKLING BOUTS: PAP Afterwards Mann-Whitney U test : PAP vs. afterwards d1 vs. d2 PAP vs .d2 day1 day2 n 119 21 32 z -3 .252 -2 .253 -1.554 n 62 15 23 z -3 .114 -3 .38 -0 .949 median 2.2% 19.1% 9 .2% p 0.0005 0.0122 n.s. median 14 min 30 min 16 min p 0.0009 0.0005 n.s. a::: :::, 0 I a::: w a.. (/) z Q (/) (/) w a::: Cl Cl 4: 100 90 80 70 60 0 f- Outliers 0 0 bd! 1 .5% spread 75% 0 Median 25% spread 50 0 -- 40 30 20 0 8 0 0 0 10 0 0 T -.- PUP YEARLING (PUP) (YEARL .) PUP+ YEARLING FEM ALE W 1TH : 8.4 Frequency of female-offspring afggressive interactions in female -pup pairs, female-yearling pairs and female-pup-yearling trios. Number of individuals included in each group are the same as in Fig. 8.4. 77 reason; aggressions would start anytime within a suckling bout, and most (73%, n=243) were not associated with female movements. This kind of aggression, unrelated to thermoregulatory movements, was never observed in single - offspring females nor directed to the pup, in the case of females suckling a pup and a yearling together. Yearlings were observed attacking their younger siblings in 11 /46 sessions (1 -12 times/session). These 'attacks' involved just open- mouth- threats and sometimes pushing the pup away from the mother. In all cases the female reacted by placing herself between the two young and threatening the yearling. DISCUSSION: Females suckling yearlings made shorter visits ashore and made longer foraging trips than mothers of recently born pups (Chapter 7). Are these differences correlated to differences in the energetic demands of pups and yearlings? Although the samples described in this chapter only cover a limited range of daylight hours and, therefore, neither the absolute proportion of time spent suckling while the mother was ashore nor the absolute frequency of suckling bouts (total number of bouts per visit) were determined, the results presented here provide indices of the energy demands of young of different ages and suggest that yearlings obtain more milk during their mothers' visit than pups. Because of their smaller body size (compared to yearlings), pups may reach satiation sooner, resulting in shorter suckling bouts. As body size increased with age, the duration of bouts and the proportion of time spent sucking increased. Yearlings, however, characte-ristically had much longer suckling bouts. It has been shown for the Galapagos fur seal that longer suckling bouts are related to a higher milk intake (Trillmich 1986, Fig. 8.5), thus, yearlings might be extracting more milk per bout than pups. 350 • • 3 00 • 2 50 • • . 200 . • • • 0) 150 • • • • ~ • • •• • - ::::! < Cl Cl < w Cl I z et:'. 0 ro (/) (L :) (L LL. 0 t- z :) 0 u >- ::::! < Cl •BORN ODEAD 100 90 80 70 60 50 40 30 20 10 0 29 9 19 29 8 18 28 8 , 8 October November December DATE 90 80 70 B 60 50 40 30 20 10 0 9 19 29 9 19 29 8 18 28 8 , 8 September October November December DATE 9. 2 Daily number of live births and dead pups on beach S3 in (A) 1984 and (8) 1985. 85 females ashore . These two were plotted against the number of dead pups counted each day, and a multiple regression analysis was carried out (Table 9.3) . In both 1984 and 1985, female numbers were significantly correlated with the daily counts of dead pups (p=0 .0001 ), but only in 1984 the number of pups born contributed to the correlation, although less significantly than the female numbers (p=0.0111). Thus, it appears as though the changes in pup mortality throughout the breeding season were mainly related to changes in female density . Fig.9 .3 shows the relationship of female numbers and daily counts of dead pups. Mortality of individually known pups Mortality of marked pups during the breeding season on beach S3 did not differ significantly between 1984 and 1985 (G=0 .08, p<0 .80, df=1), and was not different from the overall seasonal mortality for the same beach (Table 9.2b; 1984: G=0.52, p<0 .5, df=1; 1985: G=2.03, p<0 .2, df=1 ), nor from mortality of marked pups on beach N5 (S3-85 vs. N5-85: G=0.27, p<0.70, df=1 ). However, pups born and marked early and in the middle of the breeding season on beach S3 had a higher mortality than pups marked late in the season. Fig. 9.4 shows the proportion, divided by sex, of individually marked pups dying at different times during the 1984 and 1985 breeding seasons. In both years , a significantly lower proportion of the late pups died (1984: G=4.107, df=1, p=0.0427 ; 1985: Early vs. Mid: G=0.069, df=1, p=0.79 n.s.; E+M vs . Late : G=5 .605, df=1, p=0 .0179) . There were no significant differences between male and female pup mortality in 1985 (G=0.029, df=1, p=0 .866) but a significantly higher proportion of female pups died in 1984 (G= 4.107, df=1, p=0 .0427) . However, most female pups died early in the season, thus, the sex difference in 1984 might reflect the seasonal variation in mortality. In 1986, a sample of 50 females and their pups were tagged in mid- November, during the peak of pupping. Most had given birth 1-2 days earlier and were extremely aggressive. Of these, three had lost their pups two days after Table 9.3 - Multiple regression of number of pups born and female numbers on daily counts of dead pups. 1984 Parameter Intercept Pups born Number of females Value -11.463 0.083 0.059 n = 44, r2 = 0.632, p = 0.00001 1985 Parameter Intercept Pups born Number of females Value -22 .968 0.05 0.082 n = 74, r2 = 0.766, p = 0.00001 St. Err. 0.031 0.0075 St. Err. 0 .049 0.01 t-value 2.66 7.86 t-value 1.016 7.87 p 0 .0111 • 0.00001 ... p 0 .3131 n.s. 0 .00001 ... (/) a.. :J a.. Cl -- ::! - A < f- z w u 40 O::'. w 0... 20 0 EARLY MID LATE • DEAD MALES TIME DURING THE BREEDING SEASON BI DE AD FEM A LES D LIVE MALES ~ LIVE FEM AL ES 15 36 12 100 80 w 60 <.:> < f- z w u 40 O::'. w 0... B 20 0 EARLY MID LATE TIME DURING THE BREED ING SEASON 9. 4 Variations in pup mortality throughout the breeding season. Columns indicate the proportion of pups of each sex tagged early, mid and late in the breeding season that died or survived in (A) 1984 and (B) 1985 on beach S3. Sample sizes appear at the top of each column. 86 capture and only eight pups (16%) were still alive a month later. This high mortality could have bee n caused by the massive disturbance to the rookery caused by capturing 50 females in three days, but it might also reflect the very high mortality observed at that time of the year. Age at death and causes of mortality Most of the marked pups that died, did so within 15 days after marking (Fig . 9.5, Median = 9 d, n=59) . Pups which were abandoned by their mothers after capturing (3-6%, Table 9.4), died very soon afterwards. Pups in group 3 (43-45%), which were seen with their mothers at varying intervals after the perinatal attendance period, survived for longer than those in other groups, and spent a significantly lower proportion of time with their mothers than the surviving pups (F=6.368, df=1, p=0 .0254; Dead : x = 0.17±0.027 d, Surviving: x = 0.25±0.017 d; n=14 for both). Only two pups in group 3 were actually observed to be killed from attacks by other females, but some died suddenly, without signs of emaciation and thus might have been killed by aggressive females . Of the pups born to females marked in 1984 and resighted in 1985, which were not handled within the 1985 season, 6 out of 15 ( 40%) died between 14-60 days after birth (x = 18.5±3.2 d) . Except for the one which survived for 60 days, all the others died while their mothers were away at sea foraging, shortly after the end of the perinatal period. Some of the pups that were never abandoned by their mothers but died shortly afterwards (group 4), might have also died as a result of another female's attack. Most of these pups were very tame and did not try to escape when they were be ing marked . Pups which survived longer, were in general more aggressive, even shortly after birth . From the reaction of a pup while being handled, one could generally predict its likelyhood of survival. On S3 and N5, visits by sub-adult male sea lions to capture fur seal pups were common early and very late in the breeding season. Around the peak of 12 10 Med. = 9 d 8 n = 59 a: UJ Cl) ~ 6 :::, z 4 2 0 0 5 10 15 20 25 30 35 40 AGE AT DEATH (d) 9. 5 Frequency distribution of age at death of the tagged pups that died in 1984 and 1985 on beaches S3 and NS. Table 9.4 - Mean number of days after capture of death of marked pups, according to cause of death. Year Group Death cause 1984 2 Abandoned by mother a 3 Irregular attendance or female attack b ?C 1985 Capture stress 2 Abandoned by mother a 3 Irregular attendance or female attack b 4 With female all the time d Yearling suckling simultaneously Mean days after capture (± S.E.) 06.67 ± 2.88 20.11 ± 2.89 ? 00 .00 03.67 ± 0.80 16.53 ± 2.32 04.64 ± 0.57 29 .00 a - After capture was never seen again with its mother n (%) 6 (30) 9 (45) 5 (25) 20 2 (6) 6 (17) 15 (43) 11 (31) 1 (3) 35 b - After capture was seen with its mother at varying intervals and suddenly either disappeared or was seen dead. In 2 cases, the pup was known to die from an attack from another female . c - Unknown death cause, all disappeared in late Dec. while the colony was not being observed . d - After capture the pup was never separate from its mother, unknown death cause . 87 pupping, territorial male fur seals would chase male sea lions from the rookery. Early and late in the season, sea lion attacks could take place as often as 1-5 times a day, and it seemed as though it always involved the same individual as it used exactly the same route in all visits. If the sea lion managed to obtain a pup quickly during the first visit , then it did not return that day. Although the success rate of sea lion attacks seemed to be low (only 4 out of 52 observed visits resulted in a successful pup capture), and perhaps not too important as a direct cause of pup mortality, the indirect effects of the repeated disturbances might be more important . These might result in the separation of female-pup pairs and thus, starvation of the pup or an increased probability of being attacked by another female, particularly if the pup was only recently born. Females with young pups (3-30 d) which were being observed throughout a sea lion attack, rarely became separated or got back together soon afterwards. Survival outside the breeding season After the breeding season was over, mortality decreased. None of the 1984 individually known pups which were still alive by January 1985 and were observed throughout January and February, died during this period. Furthermore, at least 11 out of 15 (73%) pups which were alive by December 1984, were still alive by October 1985. However, some of these late pups had already lost their tags by January, shortly after they were tagged. It was possible to recognize them the following year only because their mothers were also tagged. Thus, the values given below reflect only minimal survival because only those pups which kept their tags until the following year (or whose mothers were tagged) could be identified . Survival to age 1 O months was significantly higher in 1985/6 (tagged Dec. 1985, resighted Oct.-Mar. 1986, 63.3%) than in 1984/5 (51%, G=7 .32, df=1, p<0.01, Table 9.5) . There was no difference in the number of male and female pups resighted each year (G=0 .01, n.s.) . A three way analysis of variance was I I Table 9.5 - Minimal survival of pups tagged in December 1984 and 1985. ---------------------------------------------·-----------------------------------------------------------------------------Year Total taooed n Decerrber Resighted by next October O/o -------------------------------·----------·-----------------------------------------------------------------------·---------------1984 1985 209 226 106 143 50.7 63.3 ------------------·------------------------------------------------------------------------------------------------------------G = 7.32, p < 0.01, df = 1 88 carried out to determine if weight at tagging, controlled for differences between sexes and between years, had any effect on survival to age 10 months (Table 9.6, Fig. 9.6). In both years, pups which were resighted the following year were significantly heavier at tagging than those that were not resighted (p==0.0001 ). Male pups were heavier than female pups in all cases (see Chapter 10). However 1985/6 pups were significantly heavier than 1984/5 pups (p==0.0001), thus, the higher survival in 1985/6 might be due to differences in pup condition. Based on the resightings in February, March and October 1986, of pups tagged in December 1985 and January 1986, survivorship of pups born during the 1985 breeding season up to one year of age was calculated (Fig. 9.7, r==0 .99, p«0.01 ). Thus, by age one year, 42% of all pups born in 1985 survived. Assuming that tag loss was similar in 1984/5 and 1985/6 and that survival followed a similar logarithmic function, survivorship of pups born in 1984 was also calculated : only 29% of all pups survived until November 1985. Some yearlings abandoned the beach during the breeding season (see Chapter 10), therefore, survival of animals older than one year could not be evaluated. Mortality during El Nif'io Between January and March 1983, during El Nino, pup mortality was extremely high. Of about 70 pups on beach S6, at least 29 (41%, not counting those washed out to sea) were seen dead on land between 22 January and 12 February (22 days). Afterwards, of 14 pups tagged between 7-10 February on beach S3, by 7 March only three (21%) were resighted . Most pups were in very poor condition (see Chapter 10) and apparently, most died of starvation after their mothers repeatedly spent long periods at sea foraging (see Chapter 7) . Only one of the three pups surviving in March 1983 was resighted during the breeding season in 1984. He was the only pup that weighed over 9 kg when captured in February. Weights for the pups that died or disappeared ranged between 4.9- 7.9 kg ., well below the average for pups in February in 1985 and 1986 (11 kg, Chapter 10). Table 9.6 - Three-way analysis of variance of weights ot resighted and non-resighted pups, tagged in 1984/5 and 1985/6, divided by sex. Source df SS MS F p ----------------------------------------------------------------------------------------------------------·-------------------Resighted (A) 1 Year (8) 1 Sex (C) 1 AxB 1 AxC 1 BxC 1 AxBxC 1 Error• 426 Total 433 A: Resighted vs. non-resighted B: 1984/5 vs . 1985/6 C: Males vs . females 22 .63 22 .63 19.48 19.48 67.04 67 .04 1.83 1.83 0.07 0.07 4.48 4.48 8 .00 8.00 487.80 1.15 611 .35 • Large error due to individual varia1ion in age at capture (0-3 months). 19 .76 0.0001 17.01 0.0001 58.55 0.0001 1.60 n.s. 0.06 n.s. 3 .91 0.0486 6.99 0.0085. • FEMALES 0 MALES Disappeared Re sighted Disappeared Resighted 1984/5 1985/6 9. 6 Mean weight in December (±S. D.)of pups tagged in 1984/5 and 1985/6 on beach S3. Sample sizes appear at base of each column. Disappeared : not resighted the following year. Resighted: resighted at least twice the following year. (/) 0::: 0 2: > 0::: :) (/) w D < 1- z w u 0::: w 0... D 1984/5 I 1985/6 100 .......... ---------------------, 80 · ....... .. .. ... .. .. ................. .... .. ......... ..... ..... ............................................ ........................ . 60 40 20 ............................................ .................................... ... ... ...... ..... ........... ......................... . 10 -+--....-~-~-~-....-~~~---.--....--,---.----, Nov. Jan . Mar . May Jul. Sep . Nov . MONTH 9. 7 Minimal survival to one year of agel of pups born in 1984 and 1985 (calculated from tag resights). 1984/5 1985/6 y = 60.l(x-0· 3159 ) R = 1.00 y = 59.l(x-0· 1493), R = 0.99 89 Mortality of pups born to females still suckling yearlings Of 15 females which were tagged in 1984 and were seen to give birth again in 1985, six were still suckling a yearling before pupping. Of the latter, two (33%) lost their pups. Four (44%) of the females which were not suckling a yearling (n=9), lost their pups. The difference was not significant (Fisher exact probability = 0.37) . Of 102 yearlings resighted in 1985 (see Chapter 9), 36 were seen suckling together with a pup. Of these pups, 21 died (58%). This mortality was not significantly higher than the overall pup mortality for the season (G=1 .62 , p<0 .30, df=1 ). Thus, it looks as though having a yearling suckling simultaneously does not significantly increase the probability of mortality of the newborn pup. However 1984 and 1985 were apparently very good years, thus perhaps in years when food availability is lower, suckling a yearling might affect pup mortality. DISCUSSION: Pup mortality during the breeding season was high and variable . In both 1984 and 1985, pup mortality rates followed the same pattern . All pups born before late October were either stillborn or died within two days. Mortality then decreased, and subsequently steadily increased until reaching a peak around late November, when female densities were highest. From this date on, mortality dropped significantly; pups born and marked late in the season (Dec.) had a higher probability of survival. Similar incidences of stillborns and/or very early mortality outside the breeding season have been reported for several other species of otariids (Zalophus, Callorhinus, Neophoca, A.pusillus, and Eumetopias : review by Odell 1972, Otaria , Majluf, unpub. data.). It has been argued that it might be related to population density, high levels of pesticides, or infections (Odell 1972), but no conclusive explanation has yet been found . Unfortunately, with the information available, it is not possible to determine what causes the abortions 90 and early mortality in the S.American fur seal at Punta San Juan. High female densities and female thermoregulatory movements appear to be the main causes of the high mortality around late November. At this time, most females ashore were in their perinatal attendance period, when they show an increase in aggression towards any approaching seal. This aggression was exacerbated by the need to reach the water's edge for thermoregulation . Between 09 :00 and 11 :00, when the sun usually strikes the rookery, females who were lying on the back parts of the beach moved towards the intertidal areas to cool off (see Chapter 6). Females in their perinatal attendance period moved, carrying their pups in their mouths and, while in transit, they were attacked by other females (Fig. 9.8) . Very often, the pup was dropped among other pups and confusions arose. At this time of the day, it was very common to observe females fighting over pups, or pups being thrown away by females who did not want them near them. Later in the season, aggression (and mortality) decreased as most of these females went to sea to forage and only a smaller portion of the female population remained on land . Of these, only a few were still in their perinatal period . For females which dropped and lost their pup within a group of females around late November, the chances of finding it again later depended on the age of the pup. If they got separated the same morning the pup was born or if the mother-pup bond was not yet firmly established , the female was more likely to take someone elses' pup. Her own pup was then likely to be killed by other females as it moved through the rookery looking for its mother or else just died of starvation if it failed to find mother. Pups of females which were also suckling a yearling at the time had an even lower probability of survival when moving through the rookery, because the yearling kept the female from getting back to the pup. The results of the study of mortality of individually known pups showed that a high proportion of the marked pups that died, did so well after the end of 9.8 Aggressive interactions between female fur seals. 91 the perinatal attendance period . However, most of the pups in this study were still in their perinatal attendance period when captured, and some might have been separated from their mothers during the critical bond formation period. If so, this might explain the lower proportion of time mothers of pups in group 3 spent with them after the perinatal period, which might have led to starvation. Most of the pups in this group died before they were 20 days old while their mothers were at sea foraging . At this age pups are still unable to move fast enough to avoid attacks; they are thus likely to be killed when moving around aggressive females. This sample might be underestimating mortality of pups younger than one day old though. High densities and crowding during the breeding season has been shown to be an important cause of mortality in other species of pinnipeds. In the Antarctic fur seal (A.gaze/la) mortality was higher at a high density than at a low density site, and was mainly due to trampling by territorial bulls and/or starvation and injuries to recently born pups (Doidge et al. 1983). In the Northern elephant seal, crowding caused the greater proportion of mortality via the "trauma- starvation" syndrome (injury, starvation and infections as a result of separation from the mother, Le Boeuf & Briggs 1977). In the grey seal, mortality was higher on a beach where seals were temporarily crowded at high tide and in storms, than in a more open beach where seals could move back in bad weather (Anderson et al. 1979) . However, in general, all these studies suggest that the proximate cause of mortality at high density sites or crowded situations is the higher probability of failure of the mother-pup bond. The results from the present study also suggest that deaths related to disturbances such as those caused by female movements, sea lion attacks, human activities or the presence of a yearling, were due to the separation of mother and pup during the bond-formation period . Early in life, if the pup becomes separated from its mother, its chances of survival are likely to be low. Thus in seals, the establishment of a firm bond between mother 92 and pup is of importance in order to ensure pup survival. In this study, pups born and marked late in the season had a higher probability of survival until the following season. In the Antarctic fur seal (Doidge et al. 1983) pups born early in the season have a higher probability of survival. Late pups were produced by younger, less experienced females and were less likely to survive . Also, pups born later in the season had a shorter lactation than those born earlier (Doidge et al. 1986) . This shorter lactation might also contribute to the lower survival of late pups in the Antarctic fur seal. The lower survival of early and mid season pups in the present study, is probably due to the fact that both had to survive through the period of high female density and aggression, whereas the late pups were born when female densities were considerably diminished and thus, the probability of becoming separated from their mothers was lower. Mortality within the first month of life in all other species of fur seals did not exceed 20% (Table 9.7). Thus, the 40% mortality reported in this study is exceptionally high. Within Punta San Juan, almost 60% of the fur seals breed on beach S3 . The amount of space available for females and their pups on this beach is limited by the need to have access to water for thermoregulation. Thus, around noon about 1,200 females are concentrated in only about one third of the total area of the beach (approximately 0.5-1 female per m2). Even though densities in other fur seal rookeries might be higher or similar to those observed on S3 (Callorhinus at St.George, M.E.Goebel, pers.comm., A.gaze/la on Bird Is., J.P.Croxall, pers.comm.), because females do not have to make daily thermoregulatory movements, females in their perinatal attendance period can stay in one place and the amount of interactions with other aggressive females are probably much lower. The very high mortality on beach S3 at Punta San Juan during the breed1ng season might therefore be the result of increasingly high female densities combined with the need to make thermoregulatory movements. The fact that early mortality did not vary between the two years covered in this Table 9. 7 - Pup mortality in different populations of fur seals ------------------------------------------------------ ------------------------------------------------------------------------------ Species Location Mortality (age range) Source ------------------------------------------------------------------------------------------------------------------------·----------- A.australis Peru 40 .0 % (0-1 months) This study A. australis Uruguay 10-20 % (0-3 months) Vaz-Ferreira 1987 A.galapagoensis Galapagos Is. 15 .0 % (0-1 months) Trillmich 1987" A. tropicalis Gough Is. 10.0 % (0-2 months) Bester 1987 A. p.pusillus S.Africa 20.0 % (0-2 months) David 1987 A.p.doriferus Australia 15.0 % (0-2 months) Shaughnessy & Warneke 1987 A.tropicalis Prince Edward Is. 10 .2 % (0-3 months) Kerley 1987 A.tropicalis French Antarctic 15.0 % (0-4 months) Roux 1987 territory A.forsteri New Zealand 20 .0 % (0-2 months) Mattlin 1987 A.gaze/la South Georgia 23.9 % (0-12 months) Payne 1977 C.ursinus Eastern Pacific 10-15 % (0-4 months) York 1987 ------------------------------------------------------------·--------------------·--------------·----------------------------------- • Mortality in normal years only. During El Nino years up to 100% of pups may die before they reach 6 months . 93 study suggests that it is independent of differences in food availability at sea. At Punta San Juan, survival after the breeding season and up to one year of age varied between the two years studied and the differences seemed to be related to differences in pup weights in December. In 1985/6, pups were heavier and had a higher survival. Lowest survival was observed in 1983, during El Nino , when pup weights were well below the average for the time of the year. Therefore, it looks as though in birth weights or early grow1h rates (both of which are proportional to the weights in December, Chapter 10) might determine survival of a years' cohort up to one year old. Trips to sea of females suckling pups in 1985/6 were significantly shorter than in 1984/5, which suggests that food was more readily available in 1985/6 (see Chapter 7) and in 1985/6 pups grew at a faster rate than in 1984/5. Also, the analysis of seats indicates that in 1984/5 adult anchovetas were not available and that fur seals had to take a wider range of prey (Chapter 4) . ln1985/6 fur seals were feeding almost exclusively on adult anchoveta. Most of the prey items consumed by fur seals in 1984/5 were smaller in size and likely to be lower in calorific content than adult anchoveta. Females may have had to catch a greater number of fish and therefore spend longer foraging times than in 1985/6 to obtain the same amount of energy. The lower pup weights and survival could therefore have been caused by a lower foraging success of females in 1984/5. Pups born in years of high food abundance, like 1985/6, might be born heavier or grow faster, and have a higher probability of surviving until the following year and perhaps of reproducing in the future. An effect of prevailing environmental conditions in the first months of life on birth weights, juvenile survival and future reproduction of different cohorts has been shown to occur in red deer (Alban et al. 1987) and in the Galapagos marine iguana (Amb/yrhynchus cristatus, Laurie 1987). 94 CHAPTER 10 - PUP GROWTH Fur seal females at Punta San Juan suckle their young under unpredictably fluctuating environmental conditions (Chapter 3) . In El Nif'lo years, when food availability was low, females had to forage at sea for longer periods (Chapter 7) and pup survival was also low (Chapter 9) . Here I compare pup weights and pup growth rates between years under varying conditions at sea to assess whether changes in prey availability to the mother at sea affect pup growth and/or condition . I also compare the growth rates of fur seal pups at Punta San Juan with those of other species of fur seals to determine whether the long lactation in this species is related to a greater maternal effort (greater relative weight at weaning). Although many S.American fur seal pups at Punta San Juan are not weaned until they are two years old (Chapter 11) , some are weaned as yearlings and thus provide a minimal weight at weaning. METHODS: Pup weights were collected between October and March on beaches S3 and N5 in 1983 1984/5 1985/6 and 1986/7 (Table 10.1) . Animals were captured either by hand or using a choker, and weighed to the nearest 200 g on a 20 kg PESOLA spring balance inside a nylon mesh bag (see Chapter 2). Two and three way analyses of variance were carried out to assess differences in pup weights between sexes, years or beaches, using the SPSS-X statistical package on the Cambridge IBM mainframe computer (ANOVA - Hierarchical approach, SPSS Inc. 1983). Growth rates were estimated from the changes in mean weights of large, random samples of pups weighed at approximately 30 d intervals . Assuming a median pupping date of 15 November(= mean median pupping date for all three Table 10.1 - Numbers of pups weighed between 1984 and 1987. Year Beach Month Males Females Total 1984/5 S3 Oct. 12 31 43 Nov. 7 3 10 Dec. 114 97 211 Feb. 48 20 68 N5 Nov. 29 24 53 Dec. 15 18 33 Feb. 27 30 57 1985/6 S3 Oct. 8 8 16 Nov. 26 22 48 Dec. 139 111 250 Jan. 73 53 126 Feb. ' 33 20 53 Mar.' 40 32 72 N5 Nov. 17 19 36 Jan . 113 88 201 Feb.' 34 32 66 Mar.' 28 27 55 1986/7 S3 Nov. 24 26 50 Dec! 104 96 200 Feb. ' 88 67 155 Mar.' 49 43 92 Total: 1,028 867 1,895 ' Weights collected by Mr. Pedro Vasquez 95 years , see Chapter 4) , pups were taken to be on averag e 30 d old in mid- December, 60 d old in mid-January, 90 d old in mid-February and 120 d old in mid-March (detailed sample sizes are given in Table 10.1). Daily growth rates (g/d) were calculated by dividing the increase in mean weight by the number of days for a given time interval (30 or 60 d). Yearly growth rates were calculated from birth weights estimated as the mean weight of pups captured before or at the median pupping date (15 Nov. see above) . Most of the pups weighed at or before the median pupping date still had a fresh umbilicus attached and thus were probably born within the previous 48 hours. These estimated birth weights were not significantly different (± 5-10 %) from those of a small sample of pups with a very fresh umbilicus(< 24 hold) . There were no significant differences between the weights of pups collected on beaches S3 and N5 in 1984/5 and 1985/6 (Table 1 O .2a) . Thus, the data from the two beaches have been combined for comparisons across years . RESULTS: Mean birth weight (from pups with a very fresh umbilicus; 1984/5, 85/6 and 86/7 combined) for males was 5.8 kg (S.D.=0.71, range 4-6 .9 kg, n=30) and for females was 5.4 kg (S.D.=0.54, range 4.2-6.6 kg, n=33) . Male pups were significantly heavier (Fig . 10.1, Table 10.2) and grew faster (8-57%, Table 10.3) than females at each age and in all three years studied. Both males and females, doubled their birth weight in about 90d (Table 10.4) . Seasonal variations in growth Seasonal differences in growth could only be assessed in 1985/6 when a sample of pups was weighed at monthly intervals between November and March. Pups grew twice as fast between January and March (60-120 d, 69 -86 g/d) as between November and January (0-60d, 34-42 g/d , Table 10.4). Sample sizes from which these growth rates were calculated are given in Table 10.3. 14 1984 12 M 10 F I- I Q 8 w )= z < w 6 I: 4 0 30 60 90 120 AGE (d) 14 1985 12 10 I- I Q 8 w )= z < w 6 I: 4 0 30 60 90 120 AGE (d) 14 1986 12 M 10 I- I Q 8 w )= z < w 6 I: 4 0 30 60 90 120 AGE (d) 10 . 1 Variations in mean weight ( ± S.E.) with age (0-120d) for male and female pups in 1984, 1985 and 1986 (for sample sizes see Table 10.4) . All weights in kg. Table 10.2 - Analysis of variance of pup weights (kg)* A - Comparison between beaches (S3 vs. N5) , between years (1984/5 vs . 1985/6) and between sexes . Source Month (age) Year Sex Beach Explained Residual Total SS 6268 .743 32.800 262 .040 0.895 6564.477 2309.731 8874 .209 df MS 4 1567.186 1 32.800 1 262.040 1 0.895 7 1390 1397 937.782 1.6612 6.352 F 943.135 19.739 157.696 0.538 564.359 B - Comparison between 1985.. and 1986 and between sexes . Source Month (age) Year Sex Explained Residual Total SS 6365.701 295.454 360.510 7021.665 2536.216 9557.881 df MS 4 1591.425 1 295.454 1 360.510 6 1170.277 1413 1.795 1419 6.736 F 886.629 164.606 200.851 651 .996 C - Comparison between between 1984 .. and 1986 and between sexes. Source Month (age) Year Sex Explained Residual Total SS 4589.360 44.976 251.634 4589.360 1412.963 6002.322 df 3 1 1 5 966 971 MS 917.872 44.976 251.634 917.872 14.463 6.182 F 627.521 30.749 172.035 627.521 p 0.0001 0.0001 0.0001 n.s . 0.0001 p 0.0001 0.0001 0.0001 0.0001 p 0.0001 0.0001 0.0001 0.0001 • SPSS-X, ANOVA (Hierarchical approach) . Because all months were not sampled in all years, factor interactions could not be calculated. " Beaches S3 and N5 combined Table 10.3 - Weights of male and female pups (0-120 d old). 1984/5* Birth** 30 d 90 d 1985/6* Birth** 30d 60d 90d 120d 1986/7*** Birth** 30d 90d 120d Males Mean S.D. n 5.86 7.02 10.88 6.20 7.44 8.69 11.06 13.64 5.84 6.39 10.75 11 .83 .77 48 1.99 129 1.52 75 1.05 51 1 .36 139 1.48 186 1.57 67 1.39 68 .98 24 1 .11 1 04 1.28 88 1.65 49 • Weights for pups on beaches S3 and N5 combined Females Mean S.D. n 5 .19 6.27 9 .84 5.70 6.71 7.84 10.02 12.09 5.18 5 .63 9 .18 9 .87 .99 58 .87 115 1.42 50 .87 49 1.03 111 1.34 141 1.40 52 1.61 59 .82 26 .88 96 1.23 67 1.70 43 .. Birth weight = mean weight of pups captured before or at the median pupping date (Nov. 15) . ... Collected on beach S3 only Table 10.4 - Absolute and daily weight gains* of male and female pups between birth, 30, 60 90 and 120 days in 1984/5, 1985/6 and 1986/7 (For sample sizes see Table 10.3) 1984/5 Weight increase: Period 0 - 30 d 30 - 90 d 1985/6 0 - 30d 30 - 60d 60 - 90 d 90 - 120 d 0 - 60d 30 - 90 d 1986/7 0 - 30 d 30 - 90 d 90 - 120 d Males absolute daily (kg) (g/d) 1.16 3 .85 1.24 1.25 2.36 2 .58 2.49 3 .61 0.54 4.36 1.08 39 64 41 42 79 86 42 60 18 73 36 • Calculated from values on Tab. 10.3 Females absolute daily (kg) (g/d) 1.09 3 .57 1 .01 1 .13 2.19 2 .06 2.14 3 .31 0.44 3 .55 0.68 36 59 34 38 73 69 36 55 15 59 23 96 Estimates of growth rates between four months and one year of age were calculated as the difference between the weight of pups in March (1 20 d old) and the weight of yearlings (360 d old) divided by the total number of days in this period (240 d) . Since yearlings weigh , on average 16.7 kg (n=17, Table 10.5) , pup weightonly increased by about 4.0 kg throughout the winter (March-Nov, mean weight in March 1985/6: M+F= 12.7 kg , n=127, Table 10.4) . This gives a daily weight gain of about 17 g/d. Fig. 10.2 shows the average growth curve for a pup (male or female) between 0-365 d. Thus, assuming a mean birth weight of 5.5 kg , as in good years such as 1985/6, by the age of 120 d old, pups can attain 75% of their yearling weight. Inter-annual differences in growth The highest mean pup weights at all ages were recorded in 1985/6 and the lowest in 1986/7 (Table 10.4) . The difference in weights between years was significant (Table 10.2) . Growth rates however, showed significant inter-annual differences only between 0-30 d ( early summer) and 90-120 d (late summer) but not between 30-90 d (mid summer). In 1986/7, early and late summer growth rates were approximately half those recorded in 1984/5 and 1985/6 (Fig. 10.3, Table 10.4) . During the 1982/3 El Nino, pup weights were only collected in February 1983 and growth rates could not be obtained. However, in the years when growth rates were obtained, mean weight of pups in February was correlated with early (0-30 d) pup growth rates (Fig. 10.4) . Therefore, February weights could be used as an index of early pup growth for the year. There was a significant correlation between mean sea surface temperatures (SST) and weight of pups in February (R2= 0.998 , n=4, p<0.01, Fig. 10.5) . During El Nino years, when SSTs were higher, early pup growth was slower that during non-El Nino or "normal" years . Inter-specific comparison of pup weights and pup growth rates Table 10.5 - Weights of yearlings captured between Oct.-Dec. in 1984/5 and 1985/6 Males Females All n 14 3 17 Mean weight 17 .3 14 .3 16 .7 S.D. 4 .12 2.91 0.98 Rarge 9.0 - 23 .5 11 .5-17.3 9.0 - 23 .5 18 16 14 ,-... C7' .:,,: 12 '--' f- I 10 Q Lu ): 8 6 4 0 Nov. 60 120 Jan . Mar. 180 240 May Jul. AGE (d) / MONTH 300 Sep. 360 Nov. 10.2 Seasonal variations in growth of a South American fur seal pup between 0- 365d of age. 100 80 A ,...._ "U '- 0\ '-' (/) w 60 t-- < Q:'. I 40 t-- ): 0 Q:'. 20 (:) 0 8-30 30-90 90-120 • 1984 INTERVAL (d) D 1985 ~ 1986 80 8 = Birth B ,...._ 60 "U '- 0\ '-' (/) w t-- 40 < Q:'. I t-- ): 0 20 Q:'. (:) 0 8-30 30-90 90-120 INTERVAL (d) 10.3 Yearly variations in growth rates of pups between birth-30d, 30-90d and 90- 120d in 1984, 1985 and 1986 for (A} male pups and (B} female pups. ,....._ -0 ' Cl' 0 4' 0 I > 0 z '-../ (./) L&J I- < ~ I I-): 0 ~ 0 40 30 20 9 -o- MALES y = - 733 .2464 + 70 .2888x, R2 = 0 .78, p<0 .05 ... FEMALES y = - 214 .6564 + 25.1022x,R2 = 0 .92, p <0 .01 • • 10 11 MEAN WEIGHT IN FEBRUARY (kg) 12 1 0.4 Relationship be1ween weights in February and early growth rates (0-30d = Nov.-Dec.} in male and female pups collected in 1984/5, 1985/6 and 1986/7. see Tab. 10.3 and 10.4 for sample sizes. O'.l w u.. ~ I- I Q w ): 0.. :::, 0.. z ~ w L 11 9 8 7 Wt . in Feb.= 16 .34 - 0.4236 SST, R2 = 0 .998, p < 0 .01 ... WEIGHT -a- SST 1983 1984 1985 YEAR 1986 1987 23 21 19 17 15 ..., ~ (/) +l u 0 '-' 1- (/) (/) z ~ w L 10.5 Annual variations in mean pup weight (M+F combined) in February (60d old) with mean sea surface temperature (SST) in February . 97 Relative birth and weaning weights in the S.American fur seal at Punta San Juan did not vary significantly when compared to other species of fur seals (Table 10.6) . In the species here compared, pups were born weighing, on average, 11 .5% of the adult female weight (S.D.=3.45%, n=14, M+F combined) and were weaned weighing on average 37.6% of the adult female weight (S.D.= 10.26%, n=14) or at about triple their birth weight (mean=3.31 times birth weight, S.D .=0 .69, n=14). These species however, differed in their duration of the lactation period (Table 10.6) . There was a significant negative correlation between the duration of lactation and average daily growth rates throughout lactation (weaning wt . - birth wt. / duration of lactation in days, R=0 .85, n=7, p<0.01, M+F combined, Fig. 10.6). Species having a short lactation grew faster. DISCUSSION: Pup growth rates in the S.American fur seal at Punta San Juan varied with sex, with season and between years. As in all other species of fur seals, males were heavier at all ages and grew faster than females (Table 10.6). These early differences apparently are part of the growth patterns leading to the marked adult sexual dimorphism. However, male pups did not appear to spend more time sucking than females (Chapter 8) . The differences in growth between sexes might be attained either by a higher rate of energy transfer to males (more milk extracted per suckling time or higher milk fat content), or by a more efficient assimilation of milk on the part of males. Studies of energetics of the Northern fur seal have shown that female pups have a higher average daily metabolic rate than males (7 .56 vs . 10.0 W kg-1 respectively) and therefore need to use a higher proportion of the milk ingested for maintenance (Costa & Gentry 1986) . With a relatively lower metabolic rate, males may be able to allocate a greater proportion of the energy obtained from milk into growth and their faster growth rates therefore are probably a result of more efficient metabolic assimilation of milk. The Table 10.6 - Comparative growth data for fur sealsa Adult female Birth weight Weight at Olerall daily growth rates Duration of 0 Lat. Species weight (kg) (kg) weaning (kg) from birth to weaning· (g/d) Lactation M F M F M F (d) ---------------------------------------------------------------------------------------------------------------------------S. American fur sealb 55 5.8 5 .1 17.3c 14.3c 31 24 365 - ? 15 S Northern fur seal 37 5 .8 5 .2 14.1 11. 7 92 72 120 60 N S. African fur seal 57 6.0 5.5 25 .0 21 .0 63 52 300 30 S Galapagos fur seal 27 3 .9 3.4 16.0 14.0 17 15 540-1080 0 New Zealand fur seal -45 3.9 3.3 14.1 12.6 34d 31d 300 45 S Antarctic fur seal 34e 5.9e 5.4e 17.oe 13_5e 92 67 90 60 S Sub-antarctic fur sea 1f 55 4.2 4.2 1s.oh 14.4h 40 31 300 52 S a - Where source not indicated, data taken from Table 15.2 in Gentry & Kooyman 1986. b - This study. c - Weaning weights unknown. Values given here correspond to weights of yearlings= minimal weaning age (see Table10.5). d - Mattlin 1978 e - Payne 1979 f - Kerley 1985 h - Tollu 197 4 • - Average growth = Wt. at weaning - Wt at birth duration of lactation (d) ,...._ 0\ C ·c "' . 98 same study showed that milk intake increased as a function of mass.There was no difference between males and females of similar body mass, but at any given age the smaller females consumed less milk than the heavier males . Taking these differences in milk intake relative to body weight into account, in the S.American fur seal at Punta San Juan overall maternal effort throughout lactation (assuming similar weaning ages, Chapter 11) should be higher in male than in female pups because males are born heavier and always require absolutely more milk. Since a larger body size confers a greater advantage on the future reproductive success of males than on that of females in polygynous , sexually dimorphic species of mammals (Glutton-Brock et al. 1981, Le Boeuf & Reiter in press), a greater investment in males will be advantageous (see Discussion - Chapter 8) . The seasonal variations in growth rates of pups at Punta San Juan appear to follow the changes in availability of food to their mothers while at sea. The highest growth rates took place during the summer, when pelagic fish were most readily available, and the lowest during the winter, when prey were scattered and at greater depths (Chapter 3). It is also possible that the slower growth in the winter is due to a change in diet as pups, aged about 8-10 months old, start to forage partially by themselves rather than to a lower foraging success of the mothers, or perhaps a combination of both. From the available information it is not yet possible to distinguish between these factors . Similar decreases in growth rates in the later stages of lactation have been described for the New Zealand (A.forsten) and Subantarctic (A.tropicalis) fur seals (Mattlin 1981, Kerley 1985), but neither of these studies could determine what caused the slower weight gains . The differences in growth rates between years were not surprising . In El Nii"lo years (1982/3 and 1986/7), when SSTs were higher, pup growth was slow. This reduced growth rate is likely to be a result of the lower food availability at sea under El Nif'lo conditions (Chapter 3) . Mothers had to spend longer times at sea 99 feeding (Chapter 7) , and probably could not provide enough milk to their pups to sustain rapid growth rates because their foraging efficiency was low (Chapter 5) . Similar changes in pup growth related to variations in availability of food at sea have been described for the Galapagos fur seal (Trillmich & Limberger 1985) , Antarctic fur seal (Croxall et al. in press), and the New Zealand fur seal (Mattlin 1983) . Average growth rates from birth to weaning in fur seals varied inversely with the duration of lactation (Fig . 10.6) and appear to be shaped by environmental constraints . In species living in sub-polar environments, pups grow relatively rapidly during the periods of food abundance and milder weather conditions , and are weaned at an early age before winter conditions set in (Peterson 1965, Doidge et al. 1986, Table 10.6) . Pups of the two sub-polar species of fur seals ( C.ursinus & A.gaze/la) have to attain in three to four months a size that will allow them to survive harsh winter conditions and in the case of the Northern fur seal, the long migration to the winter feeding grounds. For this , they need to acquire a surface/volume ratio and blubber layer that allow long periods submerged without suffering hypothermia, and a body size that allows efficient foraging (see Chapter 12). This fast growth is attained through an increased energy transfer rate , which is mainly achieved by a higher milk fat content (Trillmich & Lechner 1985, Oftedal et al. in press.) In temperate to tropical zones, relative stability in environmental conditions through seasons allows fur seals to remain in the same area throughout the year. Consequently , the period of lactation is not restricted in duration by the need to migrate. Pups can therefore remain dependent on their mothers for nourishment for periods of between 8-36 months . Growth prior to weaning in these species occurs at a much slower rate than in the sub-polar species (Gentry et al. 1986a, Mattlin 1981, Kerley 1985, Table 10.6). but does not follow a latitudinal trend. Why do these species grow at a slower rate? Given that the closely related sub-polar species do grow rapidly, there is presumably no 100 physiological limit to doing so . Explanations for slow growth on the part of temperate species could be that : 1) there is not enough food available at sea to allow females to sustain a faster pup growth. 2) there is enough food but not easily available (deeper or widely scattered), so that females have to spend longer times at sea foraging and thus the milk delivered to the pup is mainly used for maintenance during the long absences of their mothers rather than for growth . 3) high air temperatures constrain the pups in the amount of fat that can be stored as a blubber layer in that too thick a layer might cause overheating or make it necessary to dilute milk so that increased use of water for evaporative cooling is compensated by a higher intake of water through the milk. A lower fat store orintake would then result in lower growth. However, given the wide range of habitats occupied by the species of fur seals not living in sub-polar zones, it is unlikely that their lower growth has a single cause . There are insufficient data to evaluate the relative contribution of these factors . Even though fur seal species differed widely in their duration of the lactation period, weaning weights were remarkably similar: by weaning, both sexes in all species had reached about 30% (average M+F) of the adult female size (range 30-50%) or had tripled their birth weights (range 2.3-4.1 times birth weight, Table 10.6) . It has been suggested that pinniped young have to attain a minimum weight before weaning occurs (Gentry et al. 1986a, Bowen et al. 1985, Kovacs & Lavigne 1986) . Since seals have to dive in order to acquire food, and their diving efficiency varies as a function of body size (Kooyman et al. 1983b, Gentry et al. 1986), then young fur seals might have to reach a minimum body size before they can dive efficiently enough to survive without their mothers' milk. The fact that these relationships were consistent within the otariids and phocids, across a wide range of body sizes (Fig. 10.7), appears to support this idea. This will be discussed further in Chapter 12. ,, ~ 2.0 C) ~ z: <( w ): 1 .5 C) 0 ....J 2.4 2.2 I- ): 2.0 C) ~ 1.8 z: <( w ): 1.6 <=i 0 ....J 1 .4 1.2 1.0 0.5 (, • A o Phocids • Fur seals B 1 - A .ga lapagoensis 2 - A .gaze lla 3 - C .ursinus 4 - A.forsteri 5 - A.tropicalis 6 - A.australis 7 - A .pusillus 8 - P .vitulina richardsi 1.0 16 18 D 17 D LOG . BIRTH WEIGHT 2 1.5 LOG. ADULT FEM ALE WT. 9 - P .v .vitulina 1 0 - P .fascia ta 11 - P .v .conco lor 1 2 - P .caspica 13 - P .groenlandica 1 4 - H .gry pus (U .K.) 1 5 - C .cristata 1 6 - H .gry pus (Canada) 23 2 .0 D 24 3 17 - L .carcinopha9us 18 - M.schauinslandi 1 9 - E .barbatus 20 - M.leonina (M.ls.) 21 - H.lepton4x 22 - L.weddelli 23 - M .leonina (Fa lkl. ls .) 24 - M.angustirostris 1 o. 7 Relationship between weaning weight (kg) and (A) birth weight and (8) adult female weight in phocids and fur seals (Arctocephalinae). Data for phocids taken from Kovacs & Lavigne (1986). Sources for fur seals shown in Tab. 10.6. A) y = 0.05547 + 0.9297x, R2 = 0.85, p < 0.01 8) y .. • 0.1535 + 0.8161x, R2 = 0.90, p < 0.01 --,,-,. 101 CH APTER 11 - WEANI NG: Female S.American fur seals at Punta San Juan often suckle their young for more than one year and It is not rare to see 2-3 year old immatures still sucking , or females sucking two or three young of different ages (0-3 years old) at a time . In this chapter, I will address the following questions : What proportion of females continue to suckle their young for a second year? Does this proportion vary from year to year? What determines whether a yearling is weaned or not? Does suckling a yearling affect the female's subsequent breeding success? I examine these questions by looking at the resighting patterns and behaviour of the adult females and pups tagged in 1984 and 1985. METHODS: The proportion of females continuing to suckle their young into a second year was determined from the resightings between July 1985 and February 1987 of adult females and yearlings tagged in 1984 and 1985 (Table 11 .1 , Chapter 2) . For each adult female resighted a record was kept of the days she was present ashore and whether she was alone, suckling a pup, a yearling or both. For the tagged yearlings similar records were kept, noting each time they were resighted whether they were alone, with mother or with mother and a new pup. A yearling was considered to have been weaned if : 1) it was first seen sucking regularly and subsequently was either not seen again or was still around but never again seen sucking ; 2) if it was seen to visit the rookery regularly, but was never observed to suck. In the case of the tagged adult females , if they were suckl ing a yearling, the yearling was considered to have been weaned if the female continued to visit the rookery regularly, but was never again seen sucking her yearling . Table 11.1 - Dates and number of days when tags were checked (3 times/ day) in 1985,1986 and 1987. Year 1985 1986 1987 Sources : Perbd 22 Jui. - 02 Aug. 09 Sep. - 28 Dec. 05 Jan. - 16 Jan. 12 Feb. - 25 Feb. 11 Mar. - 25 Mar. 1 2 Oct. - 18 Oct. 11 Dec. - 13 Dec. 11 Jan. - 15 Jan. 18 Feb. - 21 Feb. 1 - Pamela Saunders & Ken Darcy 2 - P. Majluf 3 - Pedro Vasquez Nurrberof days 12 110 12 14 15 7 3 5 4 Total : 182 Source 1 2 2 3 3 3 3 3 3 102 RESULTS: Adult females Of 31 females tagged (with their pups) in 1984 on beach S3, 27 were resighted in 1985 and 11 (41%) were still suckling a yearling by October 1985 (Fig. 11 .1) . Of these , six (55%) gave birth during the 1985 breeding season and five either did not produce a pup or lost it very soon afterwards, so that they were never seen with a newborn pup in 1985. Of the females that gave birth in 1985, two lost their pups in less than a month. Of seven females that did not produce a pup or lost it , all continued suckling their yearlings at least until the following March (1986) . Of the females whose new pups were still alive by March, three weaned their yearlings and one kept suckling both young together, at least until March 1986. Thus, of the 11 yearlings that were sucking in October, eight (73%) continued to suck in March. Of the three yearlings that were weaned, two left the rookery and one kept coming back regularly, but was never again seen to suck, even when his mother was ashore at the time. She suckled her pup on a different site from where she used to suck the yearling before weaning . To determine whether a dependent yearling had an effect on the females' probability of producing a pup, the proportion of females pupping was compared between the females that were seen suckling a yearling and those that were never observed with a yearling between July 1985 and March 1986 (Fig. 11 .1) . No significant differences were found (Fisher exact test, P= 0.21) . Thus, it appears that suckling a yearling had no effect on the females' fertility . 1985 yearlings In order to assess whether the tagged yearlings were representative of the whole yearling population, the changes in the total number of yearlings on beach S3 (obtained from the daily censuses , Chapter 2) throughout the 1985 rr I RESIGHTED I (27) ADULT FEMALES TAGGED IN 1 984 (31) GAVE BIRTH IN 1985 (6) YEARLING WEANED ..-,-- BY MARCH (3) PUP ST ILL ALIVE IN MARCH (4) / ~ YEARLING AND PUP STILL SUCKING IN MARCH (1) / W 1TH YEARLING IN 1985 ( 11) PUP DIED BEFORE DECEMBER (2) "" \ YEARLING STILL NEVER OBSERVED SUCKING IN MARCH WITH PUP IN 1985 ----------- (7) (5) GAVE BIRTH IN / PUP ST ILL ALIVE IN MARCH (5) /"5 (9) ~ ALONE IN 1985 PUP DIED BEFORE (16) \ NEVER OBSERVED WITH PUP IN 1985 (3) DECEMBER (4) NOT RESIGHTED (4) 11 .1 Changes in reproductive status of females tagged in 1984 on beach S3 during1985/6. (numbers in parenthesis = number of females on each group). 103 breeding season were compared to the changes in the number of tagged yearlings resighted each day, and a significant correlation was found (R 2= 0.40, p=0 .0001, n= 78, Fig. 11 .2) . On average, 22% of all yearlings ashore on any given day were tagged. Of the 208 pups tagged in December 1984 106 (51 %) were still alive and were resighted as yearlings between July 1985 and March 1986. Fig. 11 .3 shows the number of tagged yearlings resighted in Jul./Aug. and the breeding season (Oct/Dec) in 1985 and throughout the summer (Jan./Mar.) in 1986. There was no significant difference between the number of male and female yearlings resighted each time. Of the 106 yearlings resighted, 52 (49%) were observed sucking continuously from July until March, 21 (20%) were observed visiting the rookery but never seen sucking, 28 (26%) were seen sucking between July and November, but were subsequently never again seen sucking or disappeared from the rookery from December onwards, and five (5%) were not seen throughout the breeding season, but were sucking on S3 in the summer (Table 11 .2). There were no significant differences in the numbers of male and female yearlings that continued sucking or disappeared after the breeding season (m:f = 34:23 and 26:23 respectively). Only 39 (37%) of the tagged yearlings were seen being nursed together with a pup and 33 (85%) of these were still sucking by March. Some may have lost their new siblings very soon after their mothers gave birth and were never observed with a pup. The proportion of yearlings whose mothers pupped reported here is therefore a minimum figure . Thus of all tagged yearlings resighted in 1985, 57 (54%) were still sucking by March 1986. In general, if the yearling had not abandoned the rookery by December, it was still sucking. Only 15% of the 73 yearlings resighted in the summer were never seen sucking. It was not possible to determine how many of these yearlings were still sucking the following breeding season, because the numbers in most of the tags used in 1984/5 were illegible after the I \ I w Cl::'. 0 I (/) <( (/) D ~ ....J Cl::'. <( w >- u... 0 Cl::'. w CD I: ::> z ....J <( I- 0 I- 260 240 220 • • • • • • • • 200 • • ••• 180 • I • • 160 • • • 140 120 100 80 60 40 20 25 30 35 40 45 50 55 NUMBER OF TAGGED YEARLINGS ASHORE 11. 2 Relationship between the number of tagged yearlings and the total number of yearlings ashore on beach S3 on any given day throughout the breeding season. y = 3.07x + 55.993, R2"' 0.404, n = 78, p < 0.001 (./) <.:> ~ ..J 100 80 ~ 60 w >- LL 0 f5 40 Ol r: :::, z 20 0 • SUCKLING 0 ALONE !2J NOT SEEN JULY/ AUG. BR .SEASON TAGGED IN 1984, RESIGHTED IN 1985/6 TOT AL YEARLINGS RESIGTHED = 106 JAN. FEB. MAR . TIME OF THE YE AR 11.3 Numbers of tagged yearlings resighted before, during and after the breeding season in 1985/6. (./) C> ~ ...J a::: 100 80 <: 60 w >- 1.J.. 0 ~ 40 CD r: ::, z: 20 0 • SUCKLING 0 ALONE l2l NOT SEEN JULY/ AUG . BR .SEASON TAGGED IN 1984, RESIGHTED IN 1985/6 TOT AL YEARLINGS RES IGTHED = 1 06 JAN . FEB. MAR . TIME OF THE YE AR 11 .3 Numbers of tagged yearlings resighted before, during and after the breeding season in 1985/6. Table 11 .2 - Status of yearlings resighted in 1985/6 (tagged in 1984 as pups) Number of pups tagged in 1984 Total pups resighted in 1985 Status in 1985/6: 1 - Sucking continuosly between July 1985 and March 1986 2 - Never observed sucking 3 - Observed sucking between Jul.-Nov. 1985 but subsequently disappeared or never again were seen sucking 4 - Not seen until after the breeding season Total: 1 +4 - Still sucking in the summer 2+3 - Presumably weaned 208 106 52 21 28 5 106 57 49 (51%) (49%) (20%) (26%) (5%) (54%) (46%) 104 first year. To compare the proportion of yearlings that continued to suck for a second year to the proportion of adult females that did not wean their yearlings, the tagged yearlings which were never seen to suck were excluded because only females which were observed suckling a yearling were included in the analysis above . Thus, of a total 85 yearlings which were observed sucking sometime between July 1985 and March 1986, 57 (67%) were still suckling by March, which is similar to the proportion of tagged females which had not weaned their yearlings by March (73%) . The proportion of yearlings which were resighted after December was significantly lower than the proportion expected from the extrapolation of the survivorship curve for pups born in 1984 (G=11 .28, df=1, p<0.001, Fig. 11.4) Thus, it looks as though the yearlings that disappeared throughout the breeding season did not die, but were probably weaned and consequently abandoned the beach. 1986 yearlings The observations of the 1986 yearlings are less detailed than those of the 1985 animals because tags were checked only for short periods which did not cover most of the breeding season (see Table 11 .1). Even so, of the 225 pups tagged in 1985, a significantly higher proportion (63 .3%, see Chapter 9) were resighted as yearlings in 1986. There was no significant difference in the number of males and females observed (m :f = 72 :71) . Only 15 (10.5%) of the resighted yearlings were never seen sucking. The remainder were observed sucking at different times between October 1986 and February 1987, and at least 82 (57.3%, m:f = 41 :41) were seen sucking in the summer (Jan-Feb) . This proportion was not significantly different from the proportion of the yearlings resighted in 1985 that were observed sucking in the summer (54%). The values for 1986/7 are minimum figures, since in January and February 1987 tags were (/) O:'. 0 2:: > O:'. :) (/) w <.:> -t: 1- z w u O:'. w Q.. 100 "TT-----------------------, 10 .................................... y = 60.0441 * xA- 0.3108 R = 1.00 ......... .. El ····························································································································f······ Nov. 84 Jan. Mar . May Jul. 85 MONTH January 1 986 Resighted after breeding season Sep. Nov. Jan . 86 Observed Expected Re sighted 73 91 G = 11 .28 d.f. = 1 p < 0.001 Disappeared 25 7 98 98 11.4 Proportion of yearlings tagged in 1984 resighted after December 1986 (open square)compared to expected survivorship curve (closed square) (see Fig .9.7). 105 checked for only a few days at a time (Table 11 .1) and the probabilijy of missing an individual was greater. DISCUSSION: Although the sample of adult females was small and the resights of the tagged yearlings varied in accuracy, the results obtained were relatively consistent : between 50-70% of the tagged pups which were resighted as yearlings continued to suck at least until March (Table 11 .3). However, given that pup survival to one year of age may vary between cohorts (Chapter 9), then the absolute number of females suckling their young for more than a year will vary between years. An inter-annual variation in the numbers of young being weaned has been shown to occur in the Galapagos fur seal (Trillmich 1986a). This will be discussed in more detail in Chapter 12. What determines whether or not a yearling continues to suck for a second year? The observations of the tagged adult females suggest that, if a female lost or did not produce a pup, she allowed the yearling to continue suckling. Most (75%) yearlings of females whose new pups were still alive by March, were weaned. Unfortunately, it was not possible to confirm this result using the larger sample sizes obtained from the tagged yearlings because not all pups that were born to the mothers of these yearlings were observed. In the Australian fur seals and in the Cape fur seals, where some yearlings suck for a second year, they usually only do so if the new pup dies (Warneke 1979, King 1983). In red deer, barren females (those that did not give birth) continue to suck their previous year calves for longer than pregnant females (Glutton-Brock, et al. 1982) . During the breeding season, pup mortality was high (40%) and independent of whether or not the mother was still suckling a yearling (Chapter 9). Presumably many of the tagged yearlings that continued to suck into the summer were allowed to do so because their mothers lost their pups or did not Table 11 .3 - Percentage of yearlings still suckling in the summer (Jan . -Mar.) in 1985/6 and 1986/7. Sarrple Adult females Yearlings 1985/6 Yearlings 1986/7 Total resighted (yeartings) 11 106 143 • Percentage excluding those yearlings never observed sucking. Still suckling in the surrvner 8 (73%) 57 (54%) (67%)* 82 (57%) (64%)* 106 give birth. In the Galapagos fur seal however, yearlings often kill their younger siblings and they generally outcompete them for milk (Trillmich 1984) . At Punta San Juan, aggressive interactions between siblings were rare (Chapter 8) and siblicide was never observed throughout this study. Perhaps the need to obtain milk from their mothers is greater in the Galapagos yearlings . At Punta San Juan, females with yearlings were as likely to produce a pup as those without. This is unlike the Galapagos fur seal, where suckling a yearling has been shown to reduce the female's probability of pupping the following year (Trillmich 1986a) . It has been shown for other species of mammals that a female's reproductive success in one year may affect her subsequent fecundity (red deer: Glutton-Brock et al. 1983, primates: Altmann et al. 1978) . The lack of a significant effect in this study might be due to the small sample sizes. Galapagos fur seals are the only other species of fur seal in which yearlings are commonly observed to continue sucking for a second year (Trillmich 1986b) . Some cases have been reported for the Cape fur seal (A.p.pusillus, Rand 1967), Australian fur seal (A.p.doriferus, Warneke 1979), and for the S.American fur seal in Uruguay (Vaz-Ferreira 1956) and in the Falkland Is . (Bonner 1984). In all other fur seals, yearlings sucking have not been reported (A.forsteri, Mattlin 1978, A.townsendii, Pierson 1978, A.philippii, Torres 1987, A.tropicalis, Bester 1981). Pups are weaned at 10-12 months old or when the new pup is born. Only in the Antarctic and Northern fur seals ( Doidge et al. 1986, Gentry & Holt 1986) all pups are weaned before they are six months old. It thus appears that the relative frequency of occurrence of yearlings continuing to suck for a second year varies inversely with latitude (Fig. 11 .5) . In the two sub-polar species, females abandon the rookeries during the winter, after weaning their young , so there is no possibility of continuing to suck after weaning occurs . None of the other more temperate fur seals migrate away from their breeding areas outside the breeding season; thus females are available to their young throughout the year, and may or may not continue suckling them for a second D ~ ...J ::><'. u ::::> (/) OFTEN 0 w > O:'. w RARE (/) CD 0 ? (/) D ~ NEVER ...J O:'. 4: w >- -10 2 • • I I I 0 10 20 1 - Galapagos FS 2 - S.American FS (Peru) 3 - S.African FS 3,4 • 5,6,7 • I 30 LATITUDE 4 - S . American FS (Ururguay) 5 - N.Zealand FS 8 • 9 10 • • I I 40 50 60 6 - Guadalupe FS 7 - Juan Fernandez FS 8 - Kerguelen FS 9 - Antarctic FS 1 0 - Northern FS 11 . 5 Relative frequency of yearlings observed suckling in fur seals . For sources see discussion. Often: > 50% of yearlings continue to suck for a second year. Rare: only few yearlings continue suckling after the first year. Not quantified. ? : Not reported for the species but possible (weaning occurs at around 300d). Never: all pups are weaned before they are 5 months old. 107 year. The higher incidence of females suckling yearlings or older immatures in the fur seals in Peru and Galapagos, may be due to the fact that both are subject to unpredictable fluctuations in food availability (Chapter 12). 108 CHAPTER 12 - DISCUSSION In this chapter I review the results reported in Chapters 4-11 in relation to other fur seals in order to assess the extent to which differences between S.American fur seals and other species relate to high environmental temperatures ashore and / or to the unpredictable fluctuatons in food availability at sea. Then I discuss the range of adaptive responses available to female fur seals in response to this environmental variability and examine the extent to which field data match theoretical expectations. SUMMARY OF RESULTS The diet and foraging behaviour of fur seals at Punta San Juan was studied from 1983 to 1986. During this period, environmental conditions varied from very poor, during the 1983 El Nino, to the favourable conditions in 1985. At Punta San Juan, fur seals breed and forage close to the most intense upwelling canter along the coast of Peru (Chapter3) . Here, fur seals feed almost exclusively on small epi- and mesopelagic fish, with apparent preference for adult anchovetas (Chapter 4) . Reports from the commercial fishery and the analysis of seats indicate that when adult anchovetas are available at sea, fur seals take few other prey. When anchovetas are scarce, they take a wider range of prey. In 1984/5 when adult anchovetas were apparently not available, fur seals had to take a larger number of smaller fish . Fur seals at Punta San Juan hunt mostly at night when anchovetas migrate close to the surface (Chapter 5) . Deep dives are energetically more costly than shallow dives (Kooyman 1981), therefore by hunting at night fur seals may reduce the costs of foraging . During El Nit'lo, when vertical migrations of fish were limited to deeper waters, fur seals probably had to spend more energy foraging because they had to make a greater number of deep dives to reach their prey. · Breeding behaviour of fur seals was studied in 1984 and 1985, both 109 favourable years . Breeding takes place during the southern spring within a discrete season, from mid-October to late December (Chapter 6). Females attend the rookery throughout the year. Males are territorial and defend portions of the beach to prevent intruders from gaining reproductive access to females . Large territorial males are present at the rookery only during the breeding season. Territories are held only in areas with access to water. Females move freely across territory boundaries and do not aggregate in stable harems. Most births and copulations occur in the wet areas of the breeding beach. Fur seals are found on land in the highest densities in November, when most females give birth ; 90% of all births take place within a period of about 40 days. Females give birth to a single pup, often while they are still suckling their previous year's offspring . Weaning age varies between 12-24 months. Of the pups born in 1984 and in 1985, approximately 40% were weaned at one year of age (Chapter 11 ). The remainder continued to suck for a second year. Apparently, it was mainly the females which lost or did not produce a pup that allowed their yearlings to continue suckling after the first year. Mating takes place 8-10 days after pupping. From then on and until the following breeding season, females alternate their time between foraging at sea and periods on land suckling their young (Chapter 7) . The duration of trips to sea and visits ashore vary with the females' reproductive status, the offsprings' demand, the females' body size and food availability at sea. Females suckling newborn pups make the shortest trips to sea (2-3 d) and show the most regular attendance patterns . Females suckling yearlings make longer trips (4-6 d) . However, females suckling a pup and a yearling simultaneously, also make short trips, similar in duration to those of females suckling a pup only. Apparently, pups less than four months old cannot survive if females are repeatedly absent for four days or longer. During El Nii'lo in 1983, when food at sea was scarce, females frequently had to stay at sea for more than 110 four days and many pups died of starvation or became severely emaciated. The longer foraging trips of females suckling yearlings appear to be related to a greater energy demand by the young. Yearlings suck for a greater proportion of the females' time ashore and make longer suckling bouts than pups (Chapter 8) . Apparently they also extract the females' fat reserves at a faster rate than pups. Visits ashore of females with yearlings are shorter in duration than those of females with pups. Larger females are able to stay ashore for longer than smaller females, probably because they have absolutely larger fat reserves . The proportion of time spent sucking and duration of suckling bouts increase with age of the pup. Pups grow at a fast rate during the spring and summer (Nov.-Mar.), when food availability at sea is normally highest (Chapter 10) . Later in the year, probably because food at sea is less available, growth of pups is slower. Variations in food availability to the mothers affect pup weights and survival of dependent young (Chapter 9), even between favourable years. In 1984/5, when adult anchovetas were scarce, pup weights and survival were lower than in 1985/6 (Chapters 9 & 10). when anchovetas were readily available. In both years there was a relationship between pup weights and survival to one year of age. Heavy pups were more likely to survive until the next year than light pups. During El Nino years pup weights and therefore, survival are further reduced. Early mortality (0-30d) appears to be independent of changes in food availability to the mothers. Despite the differences in anchoveta availability, in the two breeding seasons covered in this study about 40% of the pups born died shortly after birth (Chapter 9) . This high early mortality was due mainly to the combination of high adult female density in the breeding beaches and high levels of aggression between females when they have to move to and from the waterline to thermoregulate . After the breeding season, when most females are at sea foraging, mortality drops . COMPARISON WITH OTHER FUR SEALS I 111 In most aspects, fur seals in Peru differ little from the overall fur seal breeding pattern. They differed from most other fur seals mainly in two respects : their high early pup mortality and the extended duration of lactation . Early pup mortality In all other species of fur seals, mortality in the first months of life did not exceed 20% (Table 9.7) . In Peru , 40% of all pups born throughout the pupping season died . As stated earlier, this was due to the effects of high air temperatures and high female densities (Chapter 9) . The combination of these two factors resulted in a high probability of mother and pup becoming separated. Young pups had low chances of surviving if their mothers were unable to find them again . They either starved to death or were killed by other aggressive females. High densities or high air temperatures independently do not seem to have such an effect on pup mortality. Similarly high female densities are commonly observed in rookeries of the Northern and Antarctic fur seals (M .E.Goebel, pers.comm.,J.P .Croxall, pers.comm.). However in these two species females do not need to make daily trips to the water to thermoregulate and are able to remain in the same place throughout their perinatal attendance period. They thus have a lower probability of becoming separated from their newborn pups. In the Galapagos Islands fur seals experience even higher air temperatures than fur seals in Peru (Umberger et al. 1986) . However, densities in Galapagos fur seal rookeries are low (Trillmich 1984) . Females with young pups use the shade provided by large boulders or lava holes to avoid overheating and at the same time keep their pups away from heavy surf . Throughout the perinatal period, females with young pups stay in these resting places, which are scattered over large areas and thus keep female densities low. Only females which do not have access to shade must move to and form the water for therrnoregulation and presumably are more likely to lose their pups. 112 At Punta San Juan, on beach S3 fur seals have little access to shaded areas and rely almost exclusively on becoming wet for thermoregulation. Thus, the area available to females and their pups is limited to the areas with access to water. However, there appears to be abundant space for breeding with access to water in adjacent beaches. Why do females aggregate on beach S3 instead of spreading to other less densely populated beaches? One reason might be that because fur seals have only bred at Punta San Juan since 1974, the high densities on S3 may simply result from it being the first beach colonized, coupled with the high site fidelity typically shown by adult females and pups. In addition, S3 may be favoured because of its relative inaccessibility from land (little disturbance from humans) and because of its rocky substrate, rendering it unsuitable for sea lions, which regularly disturb the fur seal breeding rookeries when they make incursions to capture and kill pups (Chapter 10) . These disturbances appear to be less frequent on beach S3. Thus, by breeding on beach S3 disturbances by humans and sea lions might be reduced but pup mortality increased. Duration of lactation The other respect in which fur seals in Peru differed from other fur seals was in the duration of the lactation period. Apart from the Galapagos fur seal, where young are suckled for 12 - 36 months (Trillmich 1986b}, all other fur seals wean their pups at 3-12 months of age and only rarely allow them to suckle for a second year (Chapter 11 ). The time to weaning in fur seals seems to vary according to the degree of variation in food availability between seasons (Gentry et al. 1986, Oftedal et al. in press) . The Northern and Antarctic fur seals breed at high latitudes, where seasonal changes are extreme but highly predictable . These two species show the fastest growth rates and shortest lactations. In these rich environments females can maintain fast pup growth rates during the spring and summer (Gentry 113 & Kooyman 1986b) . Thus young sub-polar fur seals quickly attain a size that enables them to survive independently during harsh winter conditions and , in the case of the Northern fur seal, their winter migration (Chapter 10) . In more temperate regions seasonal variations are less marked than in sub-polar areas and although food availability may also vary with season (Chapter 3), seals can remain in the same area and suckle their young throughout the year. More temperate fur seals normally suckle their young for periods varying between 10-12 months except for the Peruvian and Galapagos fur seals which regularly do so for longer . The long lactations in the fur seals in Peru and Galapagos have been related to the unpredictable occurrences of El Nif'lo which affect both populations (Gentry et al. 1986a) . Young weaned under El Nif'lo conditions have a higher probability of mortality since they are presumably less capable divers than adults (Kooyman et al. 1983, Gentry et al. 1986a). and probably have even more difficulty obtaining food in El Nino years than do adult females (Chapter 5) . Therefore, if young are weaned when conditions at sea have improved, their probability of survival could be increased and this typically lengthens the period of lactation. However, what are the benefits (in reproductive terms) of a long lactation in favourable years? In the Galapagos fur seal it has been shown that suckling a yearling or two year old affects the females' fertility (Trillmich 1986a) . Also, pups which are being suckled together with a yearling often die because they cannot obtain enough milk (Trillmich 1986b) . The function of parental care is to maximise the quality and survival of offspring (Fuchs 1981 ). Individuals will be selected to invest in their young only as long as the costs to their own future reproduction are not greater than the benefits they derive by increasing the chances of future reproduction of the offspring (Trivers 1985) . When confronted with the situation of having several offspring simultaneously which differ in reproductive value, where reproductive value is defined as the present value of future offspring that a particular offspring represents (Pianka 1983) , a parent should allocate its available resources 114 according to the effect those resources might have on the reproductive valu e of each offspring (Schulman & Chapais 1980, Sargent & Gross 1985, Coleman et al. 1985, Mock 1987). In unpredictably variable environments, the costs and benefits to parents and offspring will change with the environment and parents should therefore adjust the amount of investment in any given offspring accordingly (Carlisle 1982, Drent & Daan 1980) . The ability to obtain enough food to fuel their metabolism and growth will determine survival and future breeding success of fur seal weaners . A number of anatomical and physiological adaptations are necessary to enable fur seals to exploit the marine environment efficiently. Short aerobic dives are energetically more · efficient than long anaerobic dives (Gentry et al. 1986). Body oxygen stores increase with body size (Kooyman 1981, Butler & Jones 1982, Gentry et al. 1986) . The greater the body size, the longer a fur seal can remain submerged without anaerobic metabolism occurring (Fig . 12.1 ). Because fur seals have to pursue their prey underwater, by maximizing the amount of time they can remain submerged they reduce the proportion of foraging time they have to spend going to and from the surface to breathe . Thus, diving efficiency will increase with body size: large seals can make longer aerobic dives than small seals (Kooyman et al. 1983). Also, seals having to remain in water for long periods while foraging might suffer heat loss because of the high thermal conductivity of water. Large size also minimizes the surface/volume ratio, which combined with a blubber layer, allows fur seals to remain in the water for long periods without suffering hypothermia (Kooyman 1981, Steiner et al. 1984) . Fur seal young may thus have to reach a minimal body size before they can forage efficiently and not suffer problems of heat loss while in the water. Minimal weaning size and how long it takes young to attain it may in turn depend on the availability of prey at sea (Fig. 12.2). If prey availability is low and as a result early growth rates of young are slow, it may take them longer to achieve ADL (min) 4 .5 4 .0 3 .5 3 .0 2.5 2 .0 I 5 Yeerl i ng 20 35 50 65 80 95 BODY SIZE ( kg) 12. 1 Relationship between body size and aerobic dive limit (ADL) : the maximum dive duration without anaerobic metabolism taking place (redrawn from Gentry et al. 1986a). 1- :::c Q w ): L H Birth weight-+--------------'------------'--------"'""----' 'w'3 'w'1 AGE (months) 12.2 Hypothetical relationship between body size and duration of lactation. young fur seals may have to attain a minimal body size before they can be safely weaned. the minimal body size will vary inversely with food availability at sea. At high food levels (H) young may be weaned at point (W1 ). At low food levels (L) young may need to attain a greater weight before they can be weaned and this may extend the duration of lactation to W2. Early growth rates may also affect the duration of lactation. Young growing fast early in life may attain the minimal body size sooner (W3). Young growing slowly may take longer to attain the same size (W2). r 115 a size that allows independent survival. Also, because prey may be at greater depths, weaners may need to attain a greater size than would be necessary in favourable years in order to forage efficiently. Conversely, if prey were readily available so that mothers could accelerate the growth of their offspring through enhanced quantity or quality of milk, young could be safely weaned at an early age. Earlier weaning of large young in favourable years has been shown to occur in the Galapagos fur seal. In 1984 some exceptionally large pups were weaned at one year of age while in less favourable years they are usually weaned when they are two years old or older (Trillmich 1986a). The effect of suckling on the yearling's reproductive value may vary with environmental conditions and with the yearling's body size (Fig. 12.3). Suckling may have a significant effect on the survival of large yearlings only if food availability at sea is very low. At high levels of food availability, large young derive little or no benefit from sucking. If the energetic costs of having to fast ashore while the mother is at sea foraging or of having to travel back and forth from the feeding grounds to the rookery to meet the mother exceed the energy they obtain from sucking, they should wean ~hemselves. Conversely, sucking might have a significant effect on the survival of yearlings if they are small in size even at high levels of food availability. The smaller in size, the higher the level of food availability at sea that is needed before they can survive without their mothers' milk. In Peru, newborn pups had a high probability of mortality if they were born before December (Chapter 9) and during El Nino, even late born pups were likely to die of starvation because their mothers had to spend longer periods at sea foraging (Chapter 7). Thus, in general, relative to yearlings, pups have a low reproductive value which varies according to conditions at sea. The effect of milk on the pup's reproductive value also varies with conditions at sea: in El Nino years extra milk may have little effect on their probability of survival whereas in favourable r PI FOOD AVAILABILITY. 12.3 The effects of parental investment (PI) on yearling survival as a function of food availability and yearling body weight. Each curve represents the probability of survival of a yearling of a given body weight (Wt). Wt increases from right to left. Increased PI will have an effect on yearling survival if food availability F. F is the minimal level of food availability necessary for a yearling to be weaned successfully and will vary as an inverse function of the yearling's weight. 116 years without milk they are unable to survive. If food at sea is readily available and if the pup is still alive after the breeding season, suckling a yearling as well as the pup might be costly because it decreases the amount of milk available to the pup and thus reduces its probability of survival (Chapter 9). If the female does not give birth in one year or if she loses her pup during the breeding season, and because she cannot produce another pup until the following year, continuing to suckle a yearling for as long as it requests milk will be advantageous as it might increase its chances of survival and future reproduction. Suckling a yearling may also affect the females' chances of producing a pup the next year (Trillmich 1986a). These effects may vary with conditions at sea though. In the Galapagos fur seal, in 'normal' non-El Nino years, only 33% of females suckling yearlings pupped the following year, but after the very favourable 1984/5 season, 60% of females which were still suckling yearlings gave birth to a new pup (F.Trillmich, pers. comm.). Presumably, in El Nino years, the costs to the females' fertility of suckling a yearling will be in general higher than in non-El Nino years. Under extreme circumstances like the 1983 El Nil'lo, the fertility of all females, even those with no dependent young may be affected. In the Galapagos fur seal, after the 1983 event, only 11% of females without young gave birth. In other years 91% of females unaccompanied by previous young produced a pup the following year (Trillmich 1986a). Food stress due to El Nino affects fur seals in Peru mainly throughout the summer, the season when food abundance is highest in favourable years (Fig. 12.4). Because food availability in winter is normally low, females may be unable to attain sufficient body condition to maintain the energetic costs of lactation and pregnancy at the same time. The high energetic cost of suckling a yearling could affect the females' condition to such an extent that implantation does not occur or abortion takes place (Boyd 1984). In contrast, suckling a yearling may have little effect on the females' fertility under favourable conditions. The abundant food >- !:: d CD -=C X d uJ ..:( Q >~ ..:( Q 0 0 Li.. 1- :c Q LLJ ): a.. => a.. • Oct. Nov. Dec. Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan. Feb. Mar. MONTH 12.4 Seasonal variations in pup weights and food availability during favourable (black squares) and El Nino (white squares) years. Horizontal bar shows the pupping season. 117 supply typical of the Peruvian upwelling system should be sufficient for females to be able to cover the costs of pregnancy and lactation. To summarize, a lengthy lactation period will be advantageous to the mother only as long as the benefits of suckling to the yearling's reproductive value exceed the costs to her reproductive success. The effects of suckling on the yearlings' reproductive value may depend on their body size and food availability at sea. A long lactation will be favourable if: 1) the young has not attained a minimal body size to allow efficient foraging . 2) If food availability is low (prey are at greater depth) and a larger body size is needed to allow independent survival. The costs to the females' reproduction will also change depending on environmental conditions and on whether or not she manages to produce a new pup and keep it alive throughout the breeding season. If food is abundant and if the pup died, the costs of suckling a yearling for a second year may be low because It may have little effect on her chances of producing a pup the following year. If the pup is alive after the breeding season and food is abundant, or if food availability is low, suckling a yearling may be costly . It may reduce the chances of survival of the pup or the females' fertility, or both. Assuming that suckling a yearling only affects the females' fertility if food availability is low and that early growth rates determine yearling size (75% of the size of yearlings may be attained by four months of age, Chapter i 0), predictions can be made of the options for females under various environmental conditions (Table i 2. i ). Individuals which have fast early growth rates and therefore are large as yearlings should be weaned under all circumstances except if the female loses her pup prior to weaning. Individuals which are small as yearlings should not be weaned unless food availability is high enough to allow them to survive independently even at a small size . The proportion of yearlings which are weaned each year should be deiermined by the sum of individual circumstances. Within a cohort, in any given year, there is wide individual variation in pup Table 12.1 - Options for weaning under different combinations of environmental conditions and reproductive status, assuming that female fertility is only affected negatively by suckling a yearling under El Nino conditions. Early growth of yearii~ Food at sea Female fertility affected Pup alive by summer Yearling weaned? Fast High No Yes Yes Fast High No No No Fast Low Yes Yes Yes Fast Low Yes No Yes Slow High No Yes Yes Slow High No No No Slow Low Yes Yes No Slow Low Yes No No j r 118 weights (Chapter 10) . Assuming such variation is maintained until the young are one year old , those individuals which have attained a size which allows independent survival under the prevailing environmental conditions may be the ones which are more likely to be weaned. The minimal weaning size will vary inversely with food availability at sea. A combination of high early growth rates and low food availability or low early growth and high food availability should then result in a similar proportion of yearlings being weaned. The weaning data for 1985 and 1986 are consistent with this suggestion (Fig. 12.5) . In 1984 pups grew slowly initially but there was abundant food at sea the following year. In 1985 pups grew fast initially and there was a mild El Nif'lo (reduced food availability) when they reached one year of age . In both years about 40% of the yearlings were weaned and the remainder continued to suck for a second year. The ones which were weaned in 1984 were apparently those whose mothers still had a live pup in the summer. The reproductive status at the time of weaning of the mothers of yearlings in 1986 is not known. Pups born in 1986 grew at a slower rate than in the two previous years. Thus, if early growth rates and food availability determine the proportion of yearlings that are weaned, then in 1987, if food availability is lower than in 1985, a smaller proportion of yearlings should continue to suck for a second year. If it is as low or lower than in 1986, no yearlings should be weaned . This will be examined in the 1987 breeding season. CONCLUSIONS The results of this study suggest that the duration of lactation in the S.American fur seal at Punta San Juan is flexible and that the major factors determinig weaning age are : food availability at sea, the yearlings' body size and whether or not the mother has a live pup in the summer. Thus, the long lactation in the Peruvian and Galapagos fur seals may indeed be related to the environmental uncertainty caused by the unpredictable occurrences of El Nif'lo (Gentry et al 1986) . I- J: Q w ): (.'.) ~ z < w ): _J < b ~ I: 85/6-L 84/5 - H 20 L-------::zg ___ --::;::,A~~...... .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 18 N J M M J S N J M M J 1984 I 1985 1986 DATE SN J MM JS N 1987 14 ,( 12 ~ I 10 4 8 6 4 2 12.5 Relationship between early growth and the proportion of yearlings weaned . Curves based on pup weight data collected between 1984-1987 (black points show mean weight ± S.D.). Extrapolations to 12 months of age calculated assuming that weights in March (when pups are four months old) = 75% of weight as yearlings. Horizontal lines indicate minimal weaning weight at high food levels (H) and mild El Nino conditions (L) . H was calculated as 1 /3 of the average adult female weight and L assumed to be higher (see Fig . 12.2 for rationale). Proportion of yearlings weaned expressed as proportion of total range shaded. Horizontal shading shows porportion of yearlings weaned in 1985 and 1986. Cross shading shows maximal proportion of yearlings expected to be weaned in the breeding season in 1987. 119 The degree of flexibility in the duration of lactation in fur seals appears to be the most adaptive feature of their matern al behaviour. Whether or not a species is able to change the amount of time it takes to raise its young to weaning seems to be mainly dependent on the seasonal changes in the environment and varies inversely with latitude (Chapter 11 ). For species inhabiting areas where they are forced to migrate in the winter (Northern and Antarctic fur seals) , the time available for raising their young is fixed and of short duration. Individuals taking longer to wean their young will have a low lifetime reproductive success because if the young have not attained a minimal weaning size within the short time interval available, they are unlikely to survive the winter (Chapter 10). Species living in areas where seasonal changes are relatively small are able to continue suckling their young for as long as it is necessary because they reside in the same place throughout the year. The degree of flexibility in the duration of lactation in these species seems to be determined by the extent of environmental uncertainty characteristic of the habitat of the species in question . Species living in areas where their food supply varies unpredictably from year to year show the greatest flexibility (Peruvian and Galapagos fur seals) . In these species, a fixed duration of lactation would probably result in a decreased reproductive success; females weaning their young early in bad years are likely to lose two young (the pup and the yearling) instead of just one (the pup) . Females are likely to be affected by at least one, and up to three El Nino events during their reproductive lives. Since they breed only once a year, losing 2-6 young would mean losing a high proportion of their lifetime reproductive output . Fur seals in Peru may be maximizing their lifetime reproductive success by being able to vary weaning age according to conditions at sea. REFERENCES: Albon,S.D., T.H.Clutton-Brock & F.E.Guinness. (1987) Early development and population dynamics in red deer. II. Density-independent effects and cohort variation. J.Anim.Ecol. 56: 69-8 1. Altmann,J. (1974) Observational study of behaviour: sampling methods . Behaviour 49 : 227-267. Altmann,J., S.A.Altmann & G.Hausfater (1978) Primate infant's effects on mother's future reproduction . Science 203 : 1028-1030 Anderson,S.S. & M.A.Fedak. (1987) Grey seal ,(Halichoerus grypus),energetics: females invest more in male offspring. J.Zool.Lond. 211, 667-679 . Anderson,S.S., J.R.Baker, J.H.Prime & A.Baird . (1979) Mortality in grey seal pups : incidence and causes. J.Zool.Lond. 189: 407-417. Antonelis Jr. ,G.A.; C.H. Fiscus & Delong, R.L. (1984) Spring and summer prey of California sea lions, (Zalophus californianus) , at San Miguel Island, California, 1978-79. Fish. Bull. 82(1) : 67- 76. Arntz,W. (1986) The two faces of El Nillo 1982-83. Meeresforsch. 31 :1-46 Arntz,W., A.Landa, & J.Tarazona (eds.) (1985) "El Nillo" Su impacto en la fauna marina. Bol. Inst. Mar Peru - Callao . Volumen extraordinario. Bailey,K.M. & D.G.Ainley. (1982) The dynamics of California sea lion predation on pacific hake. Fish.Res. 1:163-176 Barber,R.T. & F.P.Chavez. (1986) Ocean variablity in relation to living resources during the 1982-83 El Nillo. Nature 319 : 279-285 Barber,R.T. and R.L.Smith. (1981) Coastal upwelling ecosystems. In: AR.Longhurst, ed. Analysis of marine ecosystems, pp.31-68 . Academic Press . Bartholomew,G.A. Jr. & P.G.Hoel. (1953) Reproductive behavior of the Alaska fur seal, (Callorhinus ursinus). J.Mammal. 34(4) : 417-436. Bester,M.N . (1981) Seasonal changes in the population composition of the fur seal (Arctocephalus tropicalis) at Gough Island. S.Af.J.Wild.Res. 11: 49-55 . Bigg,M.A. & I.Fawcett. (1985) Two biases in diet determination of northern fur seals (Callorhinus ursinus) . In: Beddington,J.R., R.J.H.Beverton & D.M.Lavigne (eds.) Marine mammals and fisheri es London, George Allen & Unwin. Boness,D.J.; S.S.Anderson & C.R .Cox. (1982) Functions of female aggression during the pupping and mating season of grey seals , (Halichoerus grypus,Fabricius). Can.J.Zool. 60(10): 2270- 2278 . Bonner,W.N. (1981) Southern fur seals - Arctocephalus. In : Ridgway,S .H. & R.J .Harrison (eds.) Handbook of marine mammals. Vol.l : 161-208. Bonner,W.N. (1984) Lactation strategies in pinnipeds: problems for a marine mammalian group. Symp. zool. Soc. Lond. No.51 : 253-272. Bowen,W.D., O.T.Oftedal & D.J.Boness. (1985) Birth to weaning in 4 days : remarkable growth in the hooded seal (Cystophora cristata). Can.J.Zool. 63: 2841 -2846. Boyd,I.L. (1984) The relationship between body condition and the timing of implantation in pregnant Grey seals (Halichoerus grypus) . J.Zool.Lond . 203:113-123 . Butler,P .J. & D.R .Jones. (1982) The comparative physiology of diving in vertebrates. Advances in Comparative Physiology, 8:180-364. Calder,W.A. (1984) Size, function and life history. Harvard University Press, Cambridge Mass. & London England. Cane,M.A. (1983) Oceanographic events during El Nitio. Science 222: 1189-1195. Carlisle,T.R. (1982) Brood success in variable environments: implications for parental care allocation. Anim .Behav. 30 : 824-836. Chavez,F. (In Press) The annual cycle of SST along the coast of Peru. Trop.Oc.Atm.Newslet. Chirinos de Vildoso,A. & D.E.Chuman. (1968) Validez de la lectura de otolitos para determinar la edad de la Anchoveta (Engraulis ringens) . Inst.Mar.Peru.Callao - lnf. Ng 22. Clutton-Brock,T.H., F.E.Guinness & S.D.Albon . (1982) Red deer. Behaviour and ecology of two sexes . Edinburgh University Press, Edinburgh. Clutton-Brock,T.H., F.E.Guinness & S.D.Albon. (1983) The costs of reproduction to red deer hinds. J.Anim.Ecol. 52: 367-383. Clutton-Brock,T.H., M.Major, S.D.Albon & F.E.Guinness. (1987) Early development and population dynamics in red deer. I Density-dependent effects on juvenile survival. J.Anim .Ecol. 56: 53-67. Clutton-Brock,T.H., S.D.Albon & F.E.Guinness . (1981) Parental investment in male and female offspring in polygynous mammals. Nature 289: 487-489. Coleman,R.M., M.R.Gross & R.C.Sargent. (1985) Parental Investment decision rules: a test in bluegill sunfish. Behav.Ecol.Sociobiol.18:59-66. Costa, D.P. & R.L.Gentry. (1986) Free-ranging energetics of northern fur seals. In: Gentry,R.L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Cowles,T.J ., RT.Barber, & O.Guillen. (1977) Biological consequences of the 1975 El Nitio. Science 195: 285-287. Croxall,J.P. , I.Everson, G.L.Kooyman, C.Ricketts & R.W.Davis. (1985) Fur seal diving behaviour in relation to vertical distribution of krill. J.Anim.Ecol. 54:1-8. Croxall,J .P., T.S.McCann, P.A.Prince & P.Rothery. (in press) Variation in reproductive performance of seabirds and seals at South Georgia 1976-86. In: Sahrnage ,D. (ed.) Antarctic ocean and resources variability. Springer-Verlag, Berlin. Cushing,D.H. (1982) Climate and fisheries. Academic Press (London) Inc. Da Silva,J . & J.D.Neilson. (1985) Limitations of using otoliths recovered in seats to estimate prey consumption in seals . Can.J.Aquat.Sci. 42:1439-1442 Delong,R.L. (1982) Population biology of Northern fur seals at San Miguel Island, Galifornia. Ph.D.Thesis. University of California, Berkeley. Doidge,D .W., J.P.Croxall & J.R.Baker. (1983) Density-dependent pup mortality in the Antarctic fur seal (Arctocephalus gazella) at South Georgia . J.Zool.Lond. 202(3) : 449 -460 . Doidge,D.W., T .S.McCann & J.P.Croxall. (1986) Attendance behavior of Antarctic fur seals. In : Gentry,R.L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Drent,R.H . & S.Daan. (1980) The prudent parent : energetic adjustments in avian breeding. Ardea 68: 225-252 . Feldkamp,S .D. (1985) Swimming and diving in the California sea lion, (Zalophus californianus). Ph .D.Dissertation. University of Cal ifornia, San Diego. Fisher,R.A. (1930) The genetical theory of natural selection. Oxford University Press, Oxford . Fitch,J .E. & R.L.Brownell. (1971) Food habits of the Franciscana (Pontoporia blainvillei , Cetacea : Platanistidae) from South America. Bull. Mar. Sci. 21 (2) :626-636. Fonseca,T.R. (1983) El Nif'lo, fenomeno digno de estudio . Creces 83 4(5) : 24-29 . Fowler,C. W. (1987) A review of density dependence in populations of large mammals. Current M ammalogy 1 : 401-441. Fuchs,S. (1981) Consequences of premature weaning on the reproduction of mothers and offspring in laboratory mice. Z.Tierpsychol. 55: 19-32. Gentry,R.L. & G.L.Kooyman (1986b) Fur seals : maternal strategies on land and at sea. Princeton University Press. 291 p. Gentry,R.L. & G.L.Kooyman. (1986a) Methods of dive analysis. In: Gentry,R.L. & G.L.Kooyman (eds. ) Fur seals: maternal strategies on land and at sea. Princeton University Press. Gentry,R.L. & J.R.Holt (1982) Equipment and techniques for handling northern fur seals . U.S.department of Commerce, NOAA Tech. Rep., NMFS SSRF-758. Gentry,R .L. & J.R.Holt. (1986) Attendance behavior of Northern fur seals . In: Gentry,R.L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Gentry,R .L., D.P.Costa, J.P.Croxall, J.H.M.David, G.L.Kooyman, P.Majluf, T .S.McCann, & F.Trillmich . (1986) Synthesis and conclusions. In : Gentry,R.L. & G.L .Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Gentry,R.I., G .L.Kooyman & M.E.Goebel (1986b) Feeding and diving behavior of northern fur seals. In: Gentry,R .L. & G.L.Kooyman (eds.) Fur seals : maternal strategies on land and at sea. Princeton University Press. Gj0saeter,J . & K.Kawaguchi. (1980) A review of the world resources of mesopelagic fi sh. FAO Fishe ries Technical Paper NQ 193. Glantz,M .H. & J.D.Thompson (eds.) (1981) Resource management and environmental uncertainty: Lessons from coastal upwelling fisheries. N.Y.: John Wiley & Sons. 491 p. Hamilton,J.E. (1934) The southern sea lion, (Otaria byronia, De Blainville) . Discovery Reports. VOL Vllf:269-318. Hanwell, J. (1980) Atmospheric processes. London, George Allen & Unwin. 97p. Hawes,S.D. (1983) An evaluation of California sea lion scat samples as Indicators of prey importance. M.A.in Biology Thesis, S.Francisco State University. ldyll,C.P. (1973) The anchovy crisis. Sci.Am. 228(6):22-29. IMARPE. (1969) Resultados preliminares de! primer crucero de exploracion pesquera del SNP-1 6901 . lnst.Mar.Peru,Callao Serie de lnf. Especiales N° IM-39. Jordan,R. (1966) The predation of guano birds on the Peruvian anchovy (Engraulis ringens Jenyns). CALCOFI reports XI: 105-109. Jordan,R. (1971) Distribution of anchoveta (Engraulis ringens J.) in relation to the environment. lnv.Pesq. 35(1) : 113-126. Jordan,R. & A. Chirinos de Vildoso. (1965) La anchoveta (Engraulis ringens J.) conocimiento actual sobre su biologia, ecologia y pesqueria. Inst.Mar.Peru.Callao lnf. N°6. Kajimura,H . (1984) Opportunistic feeding of the Northern fur seal, (Callorhinus ursinus), in the Eastern North Pacific Ocean and Eastern Behring Sea. NOAA Tech.Rep. NMFS SSRF-779 . Kerley,G.I.H. (1985) Pup growth in the fur seals (Arctocephalus tropicalis and A.gazella) on Marion Island. J.Zool.Lond. 205 : 315-324. Kerr,R.A. (1983) Fading El Nif'lo broadening scientists' view. Science 221: 940-941 . King,J.E. (1983) Seals of the world. British Museum (Natural History) & Oxford University Press. Kooyman,G.L. (1981) Weddel seal: consumate diver. Cambridge University Press, Cambridge. Kooyman,G.L. & F.Trillmich. (1986) Diving behavior of Galapagos fur seals. In : Gentry,R.L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Kooyman,G.L. & R.L.Gentry. (1986) Diving behavior of South African fur seals. In: Gentry,R .L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Kooyman,G .L., J.O.Billups & W.D.Farwell (1983) Two recently developed recorders for monitoring diviw activity of marine birds and mammals. In: Macdonald, A.G. & I.G. Priede (eds) . Experimental biology at sea.Academic Press. 414 pp. Kooyman,G.L., M.A. Castellini; R.W. Davis & R.A. Maue. (1983) Aerobic diving limits of immature Weddel l seals. J.Comp.Physiol. 151 :171-174. Kooyman,G.L., R.W.Davis & J.P.Croxall. (1986) Diving behavior of Antarctic fur seals. In: Gentry,R.L. & G.L.Kooyman (eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Kovacs,K.M . & D.M.Lavigne. (1986) Maternal investment and neonatal growth in phocid seals . J.Anim.Ecol. 55: 1035-1051 . Laurie.A. (1987) Marine iguanas: living on the ocean margin. Oceanus 30(2) : 54-60 . LeBoeuf,B.J. & J.Reiter. (in press.) Lifetime reproductive success in Northern elephant seals. In : Glutton-Brock, T.H. (ed.) Reproductive success. Studies of selection and adaptation in contrasting breeding systems. Chicago University Press, Chicago. LeBoeuf,B .J. & K.T.Briggs . (1977) The cost of living in a seal harem. Mammalia 41 :167-195. Leible,M., E.Alveal & S.Maldonado . (1981) Catalogo de peces que habitan las aguas costeras de la bahia de concepcion y bahia de San Vicente . (lnforme de Avance) . Pont. Univ. Catolica de Chile, Sede Regional Talcahuano, Depto. Biologia y Tecnologia del Mar. Limberger,D ., F.Trillmich, G.L.Kooyman & P.Majluf. (1983) Reproductive failure of fur seals in Galapagos and Peru in 1982-83. Trop.Ocean.Atmos.Newsl. 21 :16-17. Limberger,D.; F.Trillmich, H.Biebach, & R.D.Stevenson. (1986) Temperature regulation and microhabitat choice by free ranging Galapagos fur seal pups (Arctocephalus galapagoensis). Oecologia (Berlin)69:53-59. Lindstedt,S .L. & M.S.Boyce. (1985) Seasonality, fasting endurance and body size in mammals. Am.Nat. 125: 873-878. Longhurst,A.L. (1971) The clupeoid resources of tropical seas. Oceanogr.Mar.Biol.Ann.Rev. 9 :349- 383. Majluf,P. (1985) Comportamiento del lobo fino de Sudamerica (Arctocephalus australis) en Punta San Juan, Peru, durante "El Nif'lo" 1982-83. In: Arntz, W., A.Landa, & J.Tarazona (Eds.) "El Nino" Su lmpacto en la Fauna marina. Bol. Inst. Mar Peru - Callao. Volumen extraordinario. Majluf,P. (1987) South American fur seal (Arctocephalus australis) in Peru . In: Croxall,J.P. & R.L.Gentry (eds). The status, biology and ecology of fur seals. Seattle, NOAA Tech.Rep. NMFS 51 . Majluf,P. & F.Trillmich. (1981) Distribution and abundance of sea lions (Otaria byronia) and fur seals (Arctocephalus australis) in Peru. Z.f.S/3.ugetierkunde 46(6): 384-393. Martin,P. & P.Bateson. (1986) Measuring behaviour. An introductory guide . Cambridge University Press, Cambridge. Mattlin,R.H. (1978) Population biology, thermoregulation and site preference of the New Zealand fur seal, (Arctocephalus forsteri, LESSON, 1828), on the Open Bay Islands, New Zealand. D.Phil.in Zoology Thesis. University of Canterbury, N.Z. Mattlin,R.H. (1981) Pup growth of the New Zealand fur seal (Arctocephalus forsteri) on the Open Bay Islands, New Zealand. J.Zool.Lond. 193: 305-314. McCann,T.S. (1980) Territoriality and breeding behaviour of adult male Antarctic fur seal, (Arctocephalus gaze Ila) . J.Zool.Lond. 192: 295-310. Miller,E.H. (1975a) A comparative study of facial expressions of two species of pinnipeds. Behaviour. 53: 268-284. Miller,E.H. (1975b) Social and evolutionary implications of territoriality in adult male New Zealand fur seals (Arctocephalus forsteri,Lesson, 1828), during the breeding season. Rapp.P.-v . Reun. Cons. int. Explor. Mer, 169: 170-187. Mock,D.W. (1987) Siblicide, parent-offspring conflict, and unequal parental investment by egrets and herons. Behav.Ecol.Sociobiol.20 : 247-256. Murie,D.J. & D.M.Lavigne. (1986) Interpretation of otoliths in stomach content analyses of phocid seals : quantifying fish consumption. Can.J.Zool. 64: 1152-1157. Myers, .G. (1979) Annual variation in the slope of the 142C isotherm along the equator in the Pacific ocean. Jour.Phys.Oceanog. 9: 885-891 . . ~ . . . Payne, M.R. (1979) Growth in the Antrac tic fur seal Arcto cepha lus eazell~ . J.Zool.Lond. 187:1-20. Odell,D .K. (1972) Studies on the biology of the California sea lion and the Northern elephant seal on San Nicholas Island, California. Ph.D.Dissertation, University of California, Los Angeles . Oftedal,O.T., D.J.Bonness, & R.A.Tedman. (in press.) The behavior, physiology and anatomy of lactation in the Pinnipedia. Current Mammalogy, Plenum Press, N.Y. Pauly,D. & I.Tsukayama. (1987) The Peruvian anchoveta and its upwelling ecosystem: three decades of change. ICLARM Studies and Reviews 15. Peterson,R.S. (1965) Behavior of the Northern fur seal. D.Sc.Thesis . School of Hygiene and Public Health, Johns Hopkins University. Pianka, E.R. (1983) Evolutionary ecology. Harper & Row, Publishers, New York. Pierson,M.O. (1978) A study of the population dynamics and breeding behavior of the Guadalupe fur seal, (Arctocephalus townsendi). Ph.D.Thesis, University of California, Santa Cruz. Pitcher,K.W. (1980) Stomach contents and feces as indicators of harbor seal, (Phoca vitulina), foods in the gulf of Alaska. Fish.Bull. 78(3) : 797-798 . Ponce de Leon, A. (1983) Aspectos de la reproduccion, crecimiento intrauterino y desarrollo post-natal del lobo de dos pelos Sudamericano (Arctocephalus australis, Zimmermann, 1783), en islas de la Republica Oriental del Uruguay. Tesis para optar a la Licenciatura en Oceanografia Biologica. Universidad de La Republica, Facultad de Humanidades y Ciencias, Montevideo, Uruguay. Ponce de Leon,A. (1984) Lactancia y composicion cuantitativa de la leche del lobo fino Sudamericano, (Arctocephalus australis, Zimmermann, 1783) . Anales ILPE, Montevideo Uruguay: 34- 42. Quinn,W.H. & V.T.Neal (1982) Long-term variations in the southern oscillation , El Nit'lo, and Chilean subtropical rainfall. Fish.Bull. 81 (2) : 363-374. Quinn,W.H ., T.Neal, & S.E Antunez de Mayolo (1986) El Nit"lo occurrences over the past four and a half centuries. Paper presented at the Chapman Conference on El Nit'lo, an International Symposium. Guayaquil, Ecuador. Ramage,C.S. (1986) El Nit'lo. Sci.Am. 254(6):556-661 Rand,R.W. (1967) The Cape fur seal (Arctocephalus pusillus). 3. General behaviour on land and at sea. Rep.S.Africa dept.lndust.Div. Sea Fisheries Invest. Report 89 :1-28. Reiter,J ., N.L.Stinson & B.J.Le Boeuf. (1978) Northern elephant seal development: the transition from weaning to nutritional independence. Behav.Ecol.Socibiol. 3 : 337-367. Robinson,G. & E.M.Del Pino. (1985) El Nit'lo en las Islas Galapagos: El evento de 1982-83. Publicacion de la Fundacion Charles Darwin para las Islas Galapagos, Quito, Ecuador. Ryther,J.H. (1969) Photsynthesis and fish production in the sea. Science 166 : 72 -76. Saetersdal ,G., I.Tsukayama & B.Alegre . (1965) Fluctuaciones en la abundancia aparente del stock de anchoveta en 1959-1962. Bol.lnst.Mar Peru - Callao 1 (2) :36-85. Sargent ,R.C. & M.R.Gross. (1985) . Parental Investment decision rules and the Concorde fallacy . Behav.Ecol. Socibiol. 17: 43-45. Schulman,S.R. & B.Chapals. (1980) Reproductive value and rank relations among macaque sisters. Am.Nat. 115(4) : 580-593. Siegel,S. (1956) Nonparametric statistics for the behavioral sciences. McGraw-Hill Book Company Inc. Sokal ,R.R. & F.J .Rohlf (1981) Biometry. The principles and practice of statistics in biological research . W.H.Freeman and Company, San Francisco. SPSS Inc. (1983) SPSS-X User's Guide. McGraw-Hill Book Company. Stanier,M.W., L.E.Mount & J.Bligh (1984) Energy balance and temperature regulation . Cambridge Texts in the Physiological Sciences 4. Cambridge University Press, Cambridge. Stirling,!. (1971 a & b) Studies on the behaviour of the South Australian fur seal, (Arctocephalus forsteri,Lesson) . 1) Annual cycle, postures and calls, and adult males during the breeding season. 2) Adult females and pups. Aust.J .Zool. 19 : 243-266 & 267-273 . Tollu,B. (1974) L'Otarie de l'ile Amsterdam (Arctocephalus tropicalis tropicalis) (Gray 1872) . Ph.D.Thes is. Direction des Laboratoires Scientifiques , T.A.A.F., Paris, France . Torres,D. (1987) Juan Fernandez fur seal (Arctocephalus philippii) . In: Croxall ,J.P. & R.L.Gentry (eds). The status, biology and ecology of fur seals. Seattle, NOAA Tech.Rep. NMFS 51. Trillmich, F. (1984) Natural history of the Galapagos fur seal (Arctocephalus galapagoensis, Heller) . In : Perry, R. (Ed.) Key Environments - Galapagos. Pergamon Press, Oxford : 215-223 . Trillmich, F. (1986a) Maternal investment and sex-allocation in the Galapagos fur seal, (Arctocephalus galapagoensis). Behav.Ecol.Socibiol. 19: 157-164. Trillmich,F. (1986b) Attendance behavior of Galapagos fur seals. In: Gentry,R .L. & G.L.Kooyman (eds.) Fur seals : maternal strategies on land and at sea. Princeton University Press. Trillmich,F. (1987) Galapagos fur seal (Arctocephalus galapagoensis) . In: Croxall,J .P. & R.L.Gentry (eds). The status, biology and ecology of fur seals . Seattle, NOAA Tech .Rep . NMFS 51 . Trillmich,F. & D.Limberger. (1985) Drastic effects of El Nif'lo on Galapagos pinnipeds. Oecologia 67: 19- 22. Trillmich,F. & E.Lechner. (1986) Milk of the Galapagos fur seal and sea lion, with a comparison of the milk of eared seals (Otariidae). J.Zool.Lond. 209: 271 -277. Trillmich,F. & P.Majluf. (1981) First observations on colony structure, behavior, and vocal repertoire of the South American fur seal (Arctocephalus australis, Zimmermann, 1783) in Peru . Z.f.Si:iugetierkunde 46(5) : 310-322. Trillmich,F., G.L.Kooyman, P.Majluf & M.Sanchez-Gril'ian. (1986) Attendance and diving behavior of South American fur seals during El Nil'io in 1983. In: Gentry, R.L. & G.I.Kooyman (Eds.) Fur seals: maternal strategies on land and at sea. Princeton University Press. Trivers,R .L. (1974) Parent-offspring conflict. Amer.Zoo!. 14 : 249-264. Trivers,R .L. (1985) Social evolution. The Benjamin/Cummings Publishing Company, Inc., California. Vaz-Ferreira,R. (1956) Etologia terrestre de Arctocephalus australis (ZIMMERMANN) ("Lobo fino") en las Islas Uruguayas. Montevideo, Servicio Oceanografico y de Pesca, Trabajos sobre Islas de Lobos y Lobos Marinos, N°2, 22 pp. Vaz-Ferreira ,R. (1971) Repertorio vocal aereo y sus funciones sociales en Arctocephalus australis (Zimmerman y Otaria byronia (Blainville) .. Bol.Soc.Zool.Uruguat 1 :62-65. Vill anueva,R., R.Jordan & A.Burd . (1969) lnforme sobre el estudio de comportamiento de ca rdumenes de anchoveta. lnst.Mar.Peru,Callao . Serie de lnf .Especiales NQ IM -45 . Vinogradov,M .E. (1981) Ecosystems of equatorial upwellings. In : AR.Longhurst, ed. Analysis of marine ecosystems, pp.31 -68. Academic Press . Warneke ,R.M . (1979) Australian fur seal. In: Mammals of the seas. FAO Fish.Ser. NQS.:41-44. Whittow,G.C., D.T.Matsuura & Y.C.Lyn. (1972) Temperature regulation in the California sea lion,(Zalophus californianus) . Physiol. Zoo I. 45: 68-77. Ximenez,!., L.M .Batalles, M.Lima, A.Ponce de Leon & A.Malek . (1984) Porcentaje de pref'lez en la poblacion de Arctocephalus australis en Uruguay. Anales ILPE, Montevideo Uruguay: 34-42 . Yorl<,A.E. (1987) Northern fur seal (Callorhinus ursinus): the eastern Pacific population at the Pribiloff Islands and San Miguel Island. In: Croxall,J .P. & R.L.Gentry (eds). The status, biology and ecology of fur seals. Seattle, NOAA Tech.Rep. NMFS 51. Zuta, S. and W. Urquizo. (1972) Temperatura promedio de la superficie del mar frente a la Costa Peruana. lnst.Mar.Peru,Callao Bol. 2(8) Zuta,S ., T .Rivera, & A.Bustamante. (1978) Hydrologic aspects of the main upwelling areas off Peru . In: R.Boje and M.Tomczak, eds. Upwelling ecosystems, Berlin : Springer-Verlag . UNIVERSITY LIBRARY CAMBRIDGE