Copulation and Mate Guarding in the Northern Fulmar

COPULATION AND MATE GUARDING IN THE NORTHERN FULMAR

SCOTT A. HATCH Museumof VertebrateZoology, University of California,Berkeley, California 94720 USA and AlaskaFish and WildlifeResearch Center, U.S. Fish and WildlifeService, 1011 East Tudor Road, Anchorage,Alaska 99503 USA•

ABSTRACT.--Istudied the timing and frequency of copulation in mated pairs and the occurrenceof extra-paircopulation (EPC) among Northern Fulmars (Fulmarus glacialis) for 2 yr. Copulationpeaked 24 days before laying, a few daysbefore females departed on a prelaying exodus of about 3 weeks. I estimated that females were inseminated at least 34 times each season.A total of 44 EPC attemptswas seen,9 (20%)of which apparentlyresulted in insemination.Five successful EPCs were solicitated by femalesvisiting neighboring males. Multiple copulationsduring a singlemounting were rare within pairsbut occurredin nearly half of the successfulEPCs. Both sexesvisited neighborsduring the prelayingperiod, and males employed a specialbehavioral display to gain acceptanceby unattended females.Males investedtime in nest-siteattendance during the prelaying period to guard their matesand pursueEPC. However, the occurrenceof EPC in fulmars waslargely a matter of female choice. Received29 September1986, accepted 16 February1987.

THEoccurrence and significanceof extra-pair Bjorkland and Westman 1983; Buitron 1983; copulation (EPC) in monogamousbirds has Birkhead et al. 1985). generated much interest and discussion(Glad- I attemptedto documentthe occurrenceand stone 1979; Oring 1982; Ford 1983; McKinney behavioral contextof extra-pair copulationand et al. 1983, 1984). Becausethe males of monog- mate guarding in a colonial seabird,the North- amous species typically make a large invest- ern Fulmar (Fulmarus glacialis). Fulmars are ment in the careof eggsand young, the costof among the longest-lived birds known, and fi- being cuckolded is high, as are the benefits to delity to the same mate and nest site between the successfulcuckolder. Males are expectedto years is high (Macdonald 1977, Ollason and pursueopportunities for copulationoutside the Dunnet 1978,Hatch 1985). Only one egg is laid pair bond (Trivers 1972) and to reduce as far as per clutch, and re-laying in the same season possiblethe uncertainty of paternity for the after the loss of a clutch is unknown. The sexes young they help raise. The incidence of EPC share about equally in incubation and chick- may in general be higher in colonial species rearing duties (Hatch 1985). Thus, fulmars ex- than in solitary nesters,and the threat of cuck- hibit a highly conservativesocial system. Some oldry is postulatedto be one of the principal features of the breeding biology make this disadvantagesof colonial breeding (Alexander species(and perhaps other Procellariiformes) 1974,Hoogland and Sherman 1976). particularly interesting from the standpoint of Mate guarding may be defined asany behav- copulation behavior and potential sperm com- ior by a mated male whose principal function petition. At Pacificcolonies, fulmars arrive syn- is to reduce the likelihood of encounters be- chronouslysome 6-8 weeksbefore the first eggs tween his mate and other malesduring the time are laid. There is thus a considerableprelaying when fertilization of her eggsis possible.The period for social interaction. Foraging occurs importanceof mate guarding in the monoga- over large oceanareas in boutsof severaldays, mousmale's reproductive strategy has been rec- and it is difficult for males and females to co- ognizedin a variety of species(e.g. Beecher and ordinate perfectly their attendancepatterns at Beecher 1979; Birkhead 1979, 1982; Power and the nest site. Females are receptive to copula- Doner 1980; Power et al. 1981; Werschku11982a; tion over the whole prelaying period,and sperm are storedin specialglands in the utero-vaginal (UV) region of the oviduct (Hatch 1983).Sperm Present address. remain viable in the UV glands for several

450 The Auk 104: 450-461. July 1987 July 1987] Copulationand Mate Guardingin Fulmars 451

weeks, as frequently there is no contact be- at a nest site (the host). Visitors usually arrived on tween the male and female during the "pre- the wing, but closeneighbors also moved between laying exodus" (Warham 1964), a time of con- siteson the ground. tinuousforaging immediately before egg laying In May 1980I madenotes opportunistically on ex- (lasting up to 38 days in fulmars; Hatch 1985). tra-pair visits and completed21.3 h of dedicatedob- The combination of intermittent attendance servationson one plot of about 130 nest sites (plot Q). Observationson plot Q in 1980 encompassedthe patterns,sperm storage, and a prolonged recep- whole plot, without regard to particular focal pairs. tive period in the female renders the male ful- For 3-6 h on eachof 5 daysI watchedfor interactions mar particularly susceptibleto being cuckold- amongknown individualsand recordedon audiotape ed. the characteristicsof all visits detected. In addition, I studied the timing and frequencyof copu- whenever I detecteda visit involving known breed- lation in mated pairs, both in relation to cal- ing birds during my daily rounds of other plots, I endar date and relative to the laying dates of stopped to observethe encounter to its conclusion, individual females. Social relations outside the and recorded information on duration, behavior of pair bond, including EPC,were documentedin the sexes,and the visitor'sidentity. All visitsinvolv- a seriesof extendedwatches on a group of ing breedingbirds of oppositesex were placedin one of four categoriesfor the purposeof presentation.In known individuals during 2 yr. Finally, I ex- type ! a breeding male visited an unattendedbreed- amined patternsof nest-siteattendance during ing female,in type 2 a breedingmale visiteda breed- the prelayingperiod during 6 yr for evidence ing pair, in type 3 a breeding female visited a lone of mate guarding. breedingmale, and in type 4 a breedingfemale visited a breeding pair. A standard copulation count was conducted each METHODS afternoonor evening (3-26 May). All occurrencesof mountingin 1 h were noted, including both success- The study was conductedon the Semidi Islands, ful and unsuccessfulcopulations (see below), and the Alaska (56øN, 156øW),where an estimated440,000 ful- number of pairs on the plot was recordedat the be- marsbreed (Hatch and Hatch 1983).The mainstudy ginning and end of the hour. areaon ChowletIsland contained 6-8 plotsthat were In 1981 all behavioral observations were confined usedto monitor colonyattendance from 1976 to 1981. to plot Q. Two observersworked alone or simulta- The plotscomprised 500-700 nest sites. Monitoring neously,each observing a sampleof 24-31 nest sites began 4-8 weeks before egg laying and continued for 1-9 h daily from 17 April to 26 May, for a total throughthe late chick stage in mostyears. Daily counts of 176 h. The samesample of nest siteswas observed of single birds and pairs on the plots were made be- throughout the season,but not all sitesunder obser- tween 0900 and 1600. I also monitored several hundred vation at any one time were occupied.I sampled oc- sitesindividually, noting the attendanceof adultsand cupied sitesfor 4,173 site-hours,including 1,655site- the presenceof eggsor young eachday. Bothpro- hoursfor singlesand 2,518 site-hours for pairs.Again, ceduresprovided information on nest-siteattendance the visit involving known breeding birds was the usedin the presentanalysis (see also Hatch 1985). focalunit of behavior,and the large majorityof oc- All birds usedin the study of socialinteractions currencesprobably was detected. Identity and breed- (1980-1981)were sexedby their position in copula- ing statuswere determinedas necessaryfor birds that tion, and their breedingstatus was determined by site interactedwith focal pairs. Information recorded on attachmentand egg laying. Fulmarsoccurred in a audiotapefor eachvisit includedtime, identity and wide range of color phases,and most birds also had generalbehavior of the participants,duration of the distinguishingblack marks on the culmen. Thus, a visit, occurrenceof EPC attemptsor other physical combinationof plumage differencesand bill mark- contact,and presenceor absencein the colonyof the ings provided a reliable systemfor individual iden- visitor's mate. All occurrencesof copulation in focal tification. Initially, I sketchedeach bird's bill mark- pairs during behavior watches were recorded, and ings on a templatedrawing of the head to provide a standard1-h copulationcounts were conducteddaily permanentrecord of individualidentity. After mem- in 1981. orizing the layout of nestsites and individual markers Successfulcopulations were easily distinguished (an easytask becausesite-holding fulmars were rel- from unsuccessfulattempts because the appositionof atively sedentaryon land), all observationsof behav- the birds' cloacaewas a laboredand conspicuouspro- ior were ascribedto individualsaccording to sexand cess,as was the final thrustingby the male (lasting site number. several seconds) once cloacal contact was achieved. The basic unit of behavior ! recorded was the visit, Copulation sequencesthat included cloacal contact defined as the directed approachof one bird (the and thrusting most likely resulted in insemination.I visitor)to within 0.5 m of anotherindividual or pair refer to thoseas "completedcopulations." I use the 452 SCOTTA. HATCH [Auk, Vol. 104

RESULTS 1500 Copulations(no,/pr,/h) • 0.23 0.33 0.42 Success rate - 0,63 0,69 0.64 1980 Sample size (pr,-h) - 594 438 195 Behaviorduring copulation attempts.--The behavior of fulmars during copulation was de- scribedby Macdonald(1975). Pertinent features from my own observationsare: (1) No pre- or postcopulatorydisplays ordinarily were associated with copulation in a mated pair. (2) Cop- ulation was protracted,rarely lasting less than 1 min and sometimes 6-8 min or longer. (3) Female cooperationto the extent of allowing 0,05 0.06 0.21 0,50 0.33 0,67 0,48 0.80 0.79 1.00 cloacalcontact was essentialfor successfulcop- 58 340 504 338 21 ulation. Contact was achieved as the female gradually raised and rotated her tail to meet that of the male, who appeared unable to effect contact unless he received this response. (4) Copulation always occurred on land and usually at the nest site; it was never observed on the water. (5) Failure of copulation was com- mon, and appeared to be caused by the birds being in a poor position (e.g. a cramped nest I0 20 30 I0 20 30 I0 site or uneven footing);poorly coordinatedbe- April May June havior; a low state of motivation, as somecop- ulations were initiated but left unfinished; Fiõ. 1. Patternsof prelayinõ nest-siteattendance strong onshore winds that caused the birds to in 1980and 1981and ratesof attemptedand successful copulation observedin the indicated time periods, lose their balance;or the male being distracted by a bird (usually a nonbreeder)landing close by. There was little indication that visiting birds purposely disrupted copulations in progress; term "multiple copulation" for instancesin which a male completedtwo or more successfulcopulations nonbreederscommonly landed near site-hold- in a single mounting. In those instances,each copu- ing pairsat other times,and their behavior after lation involved the full behavioral sequence, with landing was generally passivein any case. distinct episodesof cloacalcontact and thrusting. Timingand frequency of copulations.--Nest-site In this paper,the term "nonbreeder"refers to any attendancein the prelaying period was inter- bird that did not regularly occupya site in which an mittent, with synchronousvisits to land alter- egg was laid. Two categoriesof nonbreeding fulmars nating with periods of several days when no were distinguished. Unattached nonbreeders had no birds were present (Fig. 1). I witnessed the first steadypartner and no particularsite attachment. Most landfall of the seasonin 1981, when the pre- individuals in this group probably were prospecting laying period (first landing to first eggs)lasted birds that had never bred (prebreeders).They landed repeatedlyin variousplaces on the plot and, in fact, 50 days. Fulmars copulatedfrequently during accounted for most of the visiting going on at any eachpeak in attendance,and the rate increased time. By contrast,established nonbreeders were site- steadilyas the onsetof egg laying approached. holding pairsthat behavedlike breeders,except they In 1981 the standard 1-h copulationcount was producedno eggs.Some birds in this categoryprob- divided into a 0.5-h morning segment (con- ably had bred in a previous year. Of the 62 pairs ducted between 0800 and 1030) and a 0.5-h included in the 1981 focal group, 57 were breeding afternoon segment (conducted between 1300 pairs and 5 pairs were establishednonbreeders. Ex- and 1700). Altogether, 753 pairs copulated 123 ceptwhere statedotherwise, all observationson extra- times in the morning hours (0.327 copulations. pair relationsreported below involved the malesand pair Z.h-•) and 747 pairs copulated 134 times in females of the 57 breeding pairs and other known the afternoon (0.359 copulations.pa'ir '.h •). breedingbirds with which they interacted. The mean distancebetween neighboring nest sites Thus, there was no evidence that copulation on plot Q was 1.2 m (estimatedvisually). The laying frequency varied with the time of day (G = times of individual females in this study were deter- 0.678, P > 0.3). In calculatingthe total number mined to within 24 h. of copulationsper female, however, I assumed July1987] Copulationand Mate Guarding in Fulmars 453

24 h 16 (a)

1,5

1,0 •' 1,0 0,5, 0,9 (b) '• o,5

40 30 20 I0 0 Days before laying Fig. 3. Generalizedpattern of copulationtiming in relation to laying date obtainedby combining 1981

0,6 mountingand successrates (Fig. 2) and averageddata on daily prelaying attendanceby pairs in 6 yr. The two vertical scalesassume effective day lengths of 24 h and 16 h.

•.0.2 There were no copulationsfrom 10 days until 2 daysbefore egg laying, asattendance by pairs dropped to zero in that period (Fig. 2c). In fact, the interval between the last copulation and laying was usually longer than 10 days,because birdsstayed at seaan averageof 16.9days (males) or 18.7 days (females) immediately before laying (the prelaying exodus).About one-third of the pairs were together at the nest site on the day before laying, but copulation was then ex- tremely rare (pers. obs.). Days before laying The product of mounting frequencyand suc- Fig. 2. Estimatedrates of attemptedand successful cessrate (Fig. 2a, b) was the hourly rate of com- copulationin 1981relative to individual laying sched- pleted copulationsexpected on eachday of the ules.(a) Linearregression of mountingfrequency and prelaying stage.However, pairs spent lessthan day beforelaying in 57 breedingpairs on plot Q (r2 = half their time togetherbefore laying (Fig. 2c). 0.72, P < 0.001). (b) Regressionof copulationsuccess Thus,the productof mounting frequency,suc- rate (5-day means) over the same interval as above cessrate, and coincident attendance by the pair (r2 = 0.82,P < 0.001).(c) Nest-siteattendance by pairs provided an estimateof the daily rate of com- during the prelayingperiod. The proportionof pairs pletedcopulations for an individual female(Fig. at land on a given day is an estimateof the time pairs 2d). Because of the intermittent attendance of spent together at that stage.(d) Daily rate of completed copulation estimated for an individual female birds during prenesting(Fig. 1) and a high de- as the product of functions (a), (b), and (c) above. The gree of breeding synchrony in the colony (SD verticalscale assumes that copulationoccurred only of egg dates = 3.3 days), pair attendance and during a 16-h period of daylight (seetext). copulation rates both retained a strongly peri- odic pattern relative to individual laying dates. The relationshipsin Fig. 2a and 2b probably the measuredrates of copulationapplied only apply approximately to all years. Therefore, I during a 16-h period of daylight. combinedthose functions with the 6-yr average Within pairs, mounting frequencywas pos- attendancepattern of pairs from 40 to 1 day itively correlatedwith the number of days re- before laying to produce a generalized pattern maining until egg laying (Fig. 2a), as was the of copulation frequency. The result indicated rate of successfulcopulation attempts (Fig. 2b). that effective copulation peaked an average of 454 Scott A. HATCH [Auk, Vol. 104

TABLE1. Extra-pair visits observed in 1980 and their outcomes.

Duration (min) EPC b Multiple No. copula- Type observeda < 1 1-5 6-60 > 60 Attempts Success tion ½ 1 (8 • 2) 16 1 3 4 8 11 3 1 2 (8 • •2) I 1 0 0 0 1 0 -- 3 (2 • 8) 16 0 12 3 1 5 3 2 4 (2 • 82) 0 --

• Nonrandom sample of visit types (see text). bMultiple attemptsto mount during a single visit countedas 1 EPC attempt. ßMultiple copulationsduring a single mounting countedas 1 EPC.

24 days before laying (Fig. 3). Integrating this In 1981I recorded205 visitsamong breeders. daily rate over the whole prelaying period, I Twenty-sevenEPC attemptsoccurred, but only estimated that females were inseminated a mean 3 (11%) were successful(Table 2). Two of the of either 34 times (assuminga 16-h day) or 51 three successful EPCs involved the same male times (assuminga 24-h day). and female, and one was a double copulation. The occurrenceof within-pair multiple cop- Thus, one female was inseminatedup to three ulation was exceedinglylow. I saw only two times by the sameoutside male. instancesof apparently double inseminationin The 27 EPC attemptsamong breeding birds nearly 800 copulationsobserved in 1981. One in 1981 involved 13 males, 18 females, and 20 instanceof four apparent inseminationsduring different pairings of males and females. There a single mounting was noted, but the statusof were, in addition, 9 EPC attempts (1 successful) the birds (whether mated or not) was unknown. involving a breeding male and a nonbreeding Extra-paircopulation.--Thirty-three visits in- female, and 1 unsuccessfulattempt by a non- volving breeding birds were observedin 1980, breeding male to copulatewith a breeding fe- and included equal numbersof type 1 and type male. 3 visits (Table 1). EPC attempts occurred in 17 A majority of EPC attemptsoccurred during of the 33 visits (52%), and 6 EPCs (35%) were type-1visits, but the majorityof successfulEPCs successful. Three of the 6 successful EPCs were occurred between a male at his own site and a multiple copulations,involving two apparent visiting female (type 3). Ten (91%) of the 11 inseminations in each instance. The 1980 data EPC attemptsin 1981that occurredwhen a male comprised a decidedly nonrandom sample of approacheda breeding pair (type-2 visits) in- extra-pairencounters because I detectedthe oc- volved the same male. Over several days the currence of those encountersalmost solely by bird maderepeated attempts to force copulation the observationof a characteristicmale display by flying in and landing on the backof attended (describedbelow) that occurredfrequently in females(7 different individuals) in the vicinity the context of type 1 and type 3 visits. of his nest site. In all cases,these attempts were Thirteen males, 12 females, and 13 different brief and apparently futile, as the intruder was male-female pairings were involved in the 17 quickly routedby the hostpair. One other male EPC attempts. The 6 successfulEPCs included made a similar attempt in 1981, and there was 5 different pairings. One male and female cop- one instance in 1980, both with similar results. ulated successfullytwice, once at the female's The frequencyof visits and the frequencyof nest site and once at the male's site. Both of attempted EPC increasedfrom one prenesting thesewere multiple copulationsinvolving two cycle of colony attendanceto the next (Fig. 4). apparentinseminations each. Thus, one female In the final daysof the prenestingperiod, how- was inseminatedup to four timesby a breeding ever, visitsto pairs(types 2 and 4) were limited male other than her mate. In addition to the 17 by the scarcityof pairs in attendance. EPCs among breeding birds, there was 1 suc- Other behaviorsoutside the pair bond.--Obser- cessfuland 1 unsuccessfulcopulation involving vationsin both yearsindicated that interactions different nonbreeding maleswith breeding fe- among breeding birds were prevalent outside males, and 2 unsuccessfulcopulations were at- the pair bond, but did not necessarilyinvolve temptedby breedingmales with nonbreeding attempted or successfulEPC. Altogether, 40 females. (70%) of the 57 males observed and 44 (77%) of July1987] CopulationandMate Guarding inFultnars 455

2,5

1.5

6.0 (b) 4,0

2.0 j&

17-20 5-8 16-22 Apr, May May Fig.4. Ratesof visitingand extra-pair coputation (EPC) among breederson plot Q in 1981. (a) Fre- quenciesof type-I (solid circles),type-2 (open tri- angles),type-3 (solid squares), and type-4(open circles)encounters. (b) Total frequencyof all encounter types(open squares) and frequencyof EPCattempts (solidtriangles). The threetime periodson the ab- scissacorrespond to prelayingpeaks of colonyattendance illustrated in Fig. I. the 57 females acted as a visitor or host at least once, and birds from 46 different sites (81%) were involved. Visits lasted from less than 1 rain to more than 1 h (Table 2). Visitsinvolving a singlemale and female generally lasted longer than threesomes,and in a few instances of ex- tendedabsence by their mates,neighbors spent an entire day or partsof two or more daysto- gether. The incidenceof physicalcontact, such as billingor allopreening,between a hostand visitor wasrelated to the typeof visit,and occurred mostcommonly when a femalevisited a lone male at his site. Contact between a visiting female and the male of a host pair also occurred 456 SCOTTA. HATCH [Auk,Vol. 104

TABLE3. Occurrenceof unattended femalesduring then ceasedhis display and the birds sat side the last 40 daysof the prelaying period in relation by side, occasionallybilling or allopreening, to their breeding status. and displaying together to other birds that No. landed nearby. observa- Males that attempted to mount usually were tions rebuffedas the femaleside-stepped the attempt, (nest- No. Single flew from the nest site, or reared up and dis- Statusof pair days) singles No. lodged the male if he succeededin gaining the Breeding 30,564 4,919 683 13.9 normal positionfor copulation.Brief fighting Nonbreeding• 3,390 353 81 22.9 eruptedoccasionally, but the femaleapparently Establishednonbreeders only (seeMethods). P < 0.001 (G = 19.2, ! dr) for the differencebetween percentages. hadthe optionto fly at all times,and she usually did so if the encounterbecame physically ag- gressive.A male that failed in an EPC attempt frequently, whereas males that visited pairs al- usuallyresumed the EPCD and did not attempt most never engaged in such activity with the to mount again for some time, if at all. host female. Visits among breeding birds ordinarily in- A behavior pattern observedfrequently dur- volved closeneighbors. This apparentlywas not ing type-1 and type-3 visits was so nearly ex- due to a biasedsample in which the only visits clusive to those situations that I refer to the detected were those that occurred among behavior as the extra-pair courtship display neighbors.In some instancesI first recognized (EPCD). This display, performed only by the an extra-pair encounter from behaviors such as male, was characterizedby intermittent soft the EPCD already described,not knowing the cackling directed at the host female. Brief bursts location of the visitor's nest site. When the vis- of cackling(1-2 s) were spacedat fairly regular itor returnedto its own site,it almostinvariably intervalsranging from about3 to 8 s, suggesting proved to be a closeneighbor. a relative scaleof intensity in the display. Males Some females sought contacts with several sometimesperformed the EPCD almost contin- males other than their mates. There were in- uously for 10-20 min and for more than 1 h in stances in which the same female visited three extreme instances. The EPCD was seen only or four different males in the same day or sea- during the prenesting stage and proved to be son, spending up to several minutes with each a reliable behavioralcue for sexingindividuals, and engaging in behavior normally associated the only apparent one other than copulation with mated birds, suchas billing or mutual vo- itself. Only twice was a male seendirecting this cal displays. The familiarity implied by these display toward his mate. In each instance it oc- temporary associations carried over between curred, briefly and at low intensity, after the years,as similar extra-pair interactionsbetween female returned from severaldays' absence dur- known individuals were observed in 1980 and ing which the male had remained alone at the 1981. nest site. Mate guarding.--Males behaved in several In the courseof an extended type-1 encounter ways that tended to reduce the probability of there was frequently a progressionof female type-1 encounters involving their mates. First, responsestoward passiveacceptance of the out- the proportionof occupiedsites containing pairs side male's presenceat her nest site. Ifa female's was higher than the expectedvalue (on the null initial reaction to the approach of a male was hypothesisof independent attendanceby the strongly agonistic (e.g. threatened or actual sexes)on all but 2 of 105 daysin the prelaying spitting of stomach oil), relations usually ad- period (Fig. 5). The exceptionsoccurred on 22- vanced no farther, despite persistent effort by 23 May 1980, when a high proportion of lone the male. Sometimes,however, a type-1 visit males present reduced the observedproportion that began with the female avoiding her visitor of siteswith pairs. Second,the disparity in male by moving to the far side of the nest site or and female time investment followed a char- flying out and staying away for brief periods acteristicpattern with respectto a given cycle progressedto nervous bill flicking or nibbling of prelaying attendance.The occupationof sites the male's breast feathers as he continued the by pairs approached 100% (50:50 sex ratio) on EPCD, and finally to pairlike behavior.The male days of peak attendance,but the proportion of July1987] Copulationand Mate Guarding inFulmars 457

IO0 r•=IO5 8o oo Mean T

õ 6o 0,59tSE t- .9

o c). •o • 40 ' " '' Q- 0,5:

20 %•: " '

0 • 40 60 • 0.5( Ist 2nd PeakLast- I Last % Peirs Day of attendancecycle Fig. 5. Observed and expected relationships be- Fig. 6. Changes in the sex ratio among birds at tween the percentageof birds a[ the colony and the nest sites during a typical cycle of prelaying atten- percentage of occupied sites containing pairs. The dance. Included are data from 7 peaks of prelaying expected relationship assumessimilar bu[ indepen- attendanceobserved from 1977 to 1981 (seeFig. 1 and den[ a[tendance by males and females and was Hatch 1985). Only cyclesthat attained near-maximal culatedas follows: % pairs = [F/(200 - F)]100, where valuesfor the seasonand an approximatelyequal sex P = % at[endance/100 (see Coutson and Horobin 1972). ratio at their peak were used (e.g. 3-7 May and 11- The graph includes observationson 105 days with 15 May 1980, 16-20 April and 3-8 May 1981, Fig. 1). non-zero at[endanceduring pretaying periods from 1976 to 1981. The analysis is conservativebecause assumesequal amounts of pretaying nest-site atten- it steadily declined as birds departed on the danceby malesand females,whereas males actually prelaying exodus (Fig. 7a). As the laying date spend more lime on land than females(Hatch 1985). approached,the sex ratio of birds on land was increasinglybiased toward males,approaching lone males decreasedbefore the peak and in- 100% in the last 10 days (Fig. 7b). Moreover, creasedafterward (Fig. 6). That is, with respect although the occurrenceof pairs at nest sites to each episode of attendance in the prelaying was consistentlyhigher than expectedthrough- period, malestended to be the first to arrive and out the prelaying period (Fig. 5), it was rela- the last to leave. Third, males tended to go vis- tively highest at times when nest-site atten- iting only when their mateswere absentfrom dancewas low, i.e. earlier than 35 daysand later the colony, whereas females went visiting than 20 days before laying (Fig. 7c). That is, whether their mates were present or not (last femalesthat delayeddeparture on the prelaying two columns of Table 2). Considering type-1 exodus,and thosethat visited the colony in the and type-3visits, for instance,females were more 3 weeksbefore they laid, were rarely unattend- than twice aslikely asmales to go visiting when ed by their mates. their mate was present. Finally, the females of establishednonbreeding pairs were unattend- DISCUSSION ed by their partners more frequently than the femalesof breeding pairs (P < 0.001; Table 3), The incidence of EPC in fulmars was lower consistentwith the hypothesisof male nest-site than has been reported in some other colonial attendance as mate guarding. species(Bray et al. 1975, Gladstone 1979, Rob- The temporal component of sex differences erts and Kennelly 1980,Fujioka and Yamagishi in nest-site attendance varied during the pre- 1981, Werschku11982b), lower also than in ducks laying period. Relative to individual laying (McKinney et al. 1983),and possiblylower than dates, nest-site attendance was highest from in some territorial passerines(Ford 1983, Ala- about35 to 23 daysbefore egg laying, after which talo et al. 1984; but see Monnett et al. 1984). 458 SCOTTA. HATCH [Auk,Vol. 104

Pre-layincj Among Procellariiformes, EPC has been re- • 60 '"'- Exodus (a) ported only in albatrosses (Tomkins 1983, ', McKinney et al. 1984), but male prelaying at-

• 4C ,, tendancepatterns suggestive of mate guarding have been reported widely (Davis 1957;Tickell 1962, 1968; Harris 1966; Irabet 1976). Also, the behaviorof breedersvisiting and spendingtime with other breeding birds has been noted in various speciesof petrels (Richdale 1963,Tickell and Pinder 1966, Beck and Brown 1972, Imber Loo (b) 1976). Of particular interest in fulmars was the bilateral characterof relations outside the pair 0.75 ', bond, with females commonly initiating the contacts. Spermcompetition.--Copulation occurred over the whole prelaying period and peaked about o.oj; V': v 24 daysbefore egg laying. Even the earliest cop- ulationsprobably involved the transferof sperm, (c) as indicated by the presence of sperm in an oviduct collectedon 18 April 1981, 5 weeks before the first eggsappeared in the colony (Hatch 1983).Thus, fulmars exhibit two features,sperm- storage organs in the female and a prolonged receptive period, that predisposethis speciesto spermcompetition (as defined by Parker 1970). In the event a female is inseminated by more than one male, the outcome of sperm competition will depend greatly on the systemof sperm precedence,i.e. whether sperm from the first mated male predominate, sperm from the last -40 -30 -20 -I0 0 IO 20 male supersede earlier deposits, or random Days beforeor after laying sperm mixing occurs(Wade and Arnold 1980). Fig. 7. Patternsof nest-siteattendance during pre- In mammals,modes of sperm precedencemay laying and early incubation stages,combining data be highly speciesspecific (Dewsbury and Baum- collectedin 6 yr, 1976-1981:(a) overall percentageof gardner 1981,Oglesby et al. 1981). There is little birds in nest sites,(b) sex ratio among birds present comparativeinformation available for birds, but at a given stage,and (c) standardizeddifference be- experiments with domestic fowl and captive tween observed and expected numbers of pairs (3- Mallards (Anas platyrhynchos)generally have day means) on the null hypothesisof independent shown that the most recent sperm have an ad- attendanceby malesand females.Expected values are calculated as (M x F) / N, where M = number of males vantage in fertilizing eggs (e.g. Compton et al. presentduring a count,F = number of femalespres- 1978, Cheng et al. 1983). However, in an ex- ent, N = number of nest sites observed. Here, unlike periment that simulated in some important re- Fig. 5, there is no assumptionthat malesand females spectsthe situation in fulmars, the sperm of a spendequal amountsof time at the nest site. secondmale chicken(Gallus gallus) superseded that of the first only if the second male had Thus, the potential noted previously for a high continued accessto the female (Warren and Kil- incidence of cuckoldry in fulmars appeared to patrick 1929). Eggs laid more than 24 h after be largely unrealized. The existenceof special removal of the secondmale were equally likely behaviors associatedwith EPC pursuit and de- to be fertilized by sperm from either one. These fense,however, suggeststhat the threat of cuck- results seem especially relevant to the inter- oldry has been, and remains, an important fac- pretation of copulation behavior in birds that tor influencing fulmar behavior during the inseminate repeatedly and have delayed fertil- prelaying period. ization. I postulate that the usual delay of 10- July1987] Copulationand Mate Guarding in Fulmars 459

30 days between the last insemination and fer- explanationfor the prolongedreceptive period tilization in fulmarspromotes sperm mixing and and largenumber of copulationsin fulmarspos- reducesthe last-maleadvantage. tulatesan advantageto femalesof this behavior. Male reproductivestrategies.--Sperm mixing Lumpkin (1981,1983) suggested that femalesof and a reducedlast-male advantage may explain monogamousspecies may deceivetheir mates the seemingly high frequency of insemination about the timing of their fertile period by so- (34-51 times before laying) in this species,in liciting copulationwell aheadof the time when which only one egg is laid. A male that copu- effective (fertilizing) inseminationscan occur. latesso frequently with his mate may be all but The tactic is viewed as a means to elicit more assuredof paternity even if the female has en- guardingbehavior from the male,on the prem- gagedin EPC.He would, of course,also increase ise that females benefit from increased levels of the likelihoodof being the lastmale to copulate. male guarding.Critical data for evaluatingthis The relatively high incidence of multiple cop- hypothesisare lackingin my study,because the ulation during EPC suggestsoutside males at- maximumlength of the fertile periodin fulmars tempted to increase the proportion of their is unknown. Presumably,it averagesno shorter spermin the storageglands. than about 4 weeks to accommodate normal The strategyof flooding a female with sperm variability in the prelaying exodus. to increasethe likelihood of paternity hasbeen The male'sability to restrictaccess to his mate suggestedto explain the high copulation fre- clearly was limited by the female'stendency to quency observedin the White Ibis (Eudocimus associatewith other males away from the nest aIbus; Benshoof and Thornhill 1979). Indeed, site. Inasmuch as 5 of 9 successful EPCs occurred there may be a correlationbetween copulation during type-3visits (females visiting males),and frequency and the relative threat or conse- successfulEPC appeared to require female co- quencesof cuckoldry,but few dataare available operationin any case,relations outside the pair on the frequency or total number of insemi- bond appearedto be largely a matter of female nations per reproductive cycle in wild birds. choice. The view that most EPCs in monoga- The range appearsto be once (as in certain lek- mousbirds are forcedon unwilling femaleshas king species;Wiley 1973,Oring 1982)to dozens generally prevailed (Gladstone1979, McKinney of times (Brown 1967, Burger 1976, Gochfeld et al. 1984),and the phenomenon I observedof 1980, this study). mated femalessoliciting copulationfrom out- The abundanceof sperm introduced into the side maleshas rarely been reported (McKinney female reproductivetract may thus be one ele- et al. 1983). The behavior runs counter to most ment of the male strategy,but in addition males theoretical treatments of female reproductive investedmore time in prelaying nest-siteatten- tacticsin monogamousspecies (Trivers 1972, dance than females,which I interpret as mate Gladstone 1979). guarding. It can scarcelybe viewed as site de- There are at least three ways a female might fense as there was little fighting or other ago- benefit by copulatingwith malesother than her nistic behavior except in the context of EPC. mate. First, she may confer on her offspring the Other birds rarely spent time in unoccupied genesof an especiallyfit male whose quality nest sites, even in the most crowded portions she has been able to assess(Trivers 1972). Sec- of the colony. Macdonald (1980) also noted the ond, she increasesthe lifetime genotypicvari- apparent lack of necessityfor strong nest-site ability of her offspring and therefore possibly defensein a colony of fulmars. her own fitness (Williams 1975). Third, the fe- Mate guarding and the pursuit of extra-pair male whose mate is derelict in his prelaying copulationsmay be mutually exclusivemale ac- attendance and copulation in the current sea- tivities (Beecher and Beecher 1979, Mineau and son,or that hasexperienced infertility with her Cooke 1979, Barash 1981, Werschkul 1982b). This partner in the past, may attempt to prevent its was only partially true for fulmars;by spending recurrenceby copulating with other males. time in or near his nest site, a male effectively The criteria involved in female choice in this guarded againstthe unaccompaniedreturn to system could include the attendance patterns land of his matebut wasalso free to visit neigh- of a neighboring male as an indicator of his boring femalesor be visited by them. individual quality. I haveattempted, with some Femalereproductive strategies.--An alternative success,to relatethese patterns to breedingsuc- 460 SCOTTA. HATCH [Auk, Vol. 104 cess (Hatch 1985), and female fulmars poten- ß 1982. Timing and duration of mate guarding tially have much more informationon this point in MagpiesPica pica. Anim. Behav.30: 277-283. than I. More directly, femaleshave information ß S. D. JOHNSON,& D. N. NETTLESHIP. 1985. on their neighbors' recordsof successin raising Extra-pair matings and mate guarding in the CommonMurre Uria aalge.Anim. Behav.33: 608- young, and on their survival from one breeding 619. season to the next. In fact, the combined effects BJORKLAND,g., & B. WESTMAN. 1983. Extra-pair cop- of adult mortality and changesof mate or nest ulations in the Pied Flycatcher (Ficedulahypo- site resulted in a turnover among neighborsof leuca). Behav. Ecol. Sociobiol. 13: 271-275. only about 6% per year (Hatch 1985).Thus, the BRAY,O. E., J. J. KENNELLYß& J. L. GUAmNO. 1975. wide range of female responsesto the approach Fertility of eggs producedon territories of va- of outsidemales may in part reflect the number sectomizedRed-winged Blackbirds.Wilson Bull. of yearsparticular neighbors had interacted,and 87: 187-195. instancesof successfulEPC may representa cul- BROWN,R. G.B. 1967. Courtship behaviour in the mination of several years' effort by a breeding LesserBlack-backed Gull, Larusfuscus. Behaviour 29: 122-153. male to establishfamiliarity and the required BUITRON,D. 1983. Extra-paircourtship in Black-billed level of acceptanceby a neighboring female. Magpies.Anim. Behav.31: 211-220. BURGER,J. 1976. Daily and seasonalactivity patterns ACKNOWLEDGMENTS in breeding Laughing Gulls. Auk 93: 308-323. CHENG, K. M., J. T. BURNS, & F. MCKINNEY. 1983. Early fieldwork on the Semidiswas supportedin Forcedcopulation in captive Mallards. III. Sperm part by the AlaskaOuter ContinentalShelf Environ- competition.Auk 100: 302-310. mental AssessmentProgram (U.S. Bureau of Land COMPTON,M. M., H. P. VAN KREY,& P. B. SIEGEL. 1978. Managementand National Oceanicand Atmospheric The filling and emptying of the uterovaginal Administration). My wife, Martha, contributed im- sperm-hostglands in the domestichen. Poultry portant observationsof visiting behavior and assisted Sci. 57: 1696-1700. in countlessother ways. The manuscriptbenefited COULSON,J. C., & J. g. HOROBIN. 1972. The annual from the reviews of F. A. Pitelka, R. K. Colwell, M. L. re-occupationof breeding sites by the Fulmar. Morrison, T. R. Birkhead,A. R. DeGange,N. L. Ford, Ibis 114: 30-42. and M. P. Lombardo.Most of all I expressmy appre- DAVIS,P. 1957. The breeding of the Storm Petrel. ciation to Frank Pitelka, whose teaching and enthu- Brit. Birds 50: 85-101,371-384. siasmfor the subjecthave greatly increasedmy own DEWSBURY,D. A., & D. J. BAUMGARDNER.1981. Stud- awarenessand enjoymentof avian socialsystems. This iesof spermcompetition in two speciesof muroid paper is dedicatedto him on the occasionof his 70th rodents. Behav. Ecol. Sociobiol. 9: 121-133. birthday. FORD,N. L. 1983. Variation in mate fidelity in monogamousbirds. Pp. 329-356 in Current orni- LITERATURE CITED thology, vol. 1 (R. F. Johnston,Ed.). New York, Plenum Press. ALATALO, R. V., L. GUSTAFSSON,& A. LUNDBERG. 1984. FUJIOKA,M., & S. YAMAGISHI. 1981. Extramarital and High frequencyof cuckoldryin Pied and Col- pair copulationsin the Cattle Egret.Auk 98: 134- lared flycatchers.Oikos 42: 41-47. 144. ALEXANDER,R. D. 1974. The evolution of social be- GLADSTONE,D.E. 1979. Promiscuityin monogamous havior. Ann. Rev. Ecol. Syst.5: 325-383. colonial birds. Amer. Natur. 114: 545-557. BARASH,D.P. 1981. Mate guardingand gallivanting GOCHFELD,M. 1980. Mechanismsand adaptive value by male hoary marmots(Marmota caligata). Behav. of reproductive synchrony in colonial seabirds. Ecol. Sociobiol. 9: 187-193. Pp. 207-270 in Behavior of marine animals. Vol. BECK,J. R., & D. W. BROWN.1972. The biology of 4, Marine birds (J. Burger, B. L. Olla, and H. E. Wilson's Storm Petrelß Oceanitesoceanicus (Kuhl), Winn, Eds.). New York, Plenum Press. at Signy Island,South Orkney Islands.Brit. Ant- HARMS,M.P. 1966. Breedingbiology of the Manx arctic Surv. Sci. Rept. No. 69. ShearwaterPuffinus puffinus. Ibis 108: 17-33. BEECHER,M.D., & I. M. BEECHER.1979. Sociobiology HATCH,S.A. 1983. Mechanism and ecologicalsig- of Bank Swallows: reproductive strategy of the nificanceof sperm storagein the Northern Ful- male. Science 205: 1282-1285. mar with reference to its occurrence in other birds. BENSHOOFßL., & R. THORNHILL. 1979. The evolution Auk 100: 593-600. of monogamy and concealedovulation in hu- 1985. Population dynamics,breeding ecol- mans. J. Soc. Biol. Struct. 2: 95-106. ogy, and socialbehavior of the Northern Fulmar BIRKHEAD,T. R. 1979. Mate guarding in the Magpie (Fulmarusglacialis). Ph.D. dissertation,Berkeley, Pica pica.Anim. Behav. 27: 866-874. Univ. California. July1987] Copulationand Mate Guarding in Fulmars 461

ß& M. A. HATCH. 1983. Populations and hab- lutionary consequencesin the insects.Biol. Rev. itat use of marine birds in the Semidi Islands, 45: 525-567. Alaska. Murrelet 64: 39-46. POWER,H. W., & C. G. P. DONER. 1980. Experiments HOOGLAND,J. L., & P. W. SHERMAN. 1976. Advan- on cuckoldry in the Mountain Bluebird. Amer. tagesand disadvantagesof Bank Swallow (Riparia Natur. 116: 689-704. riparia)coloniality. Ecol. Monogr. 46: 33-58. ß E. LITOVICH, & M.P. LOMBARDO. 1981. Male IMBER,M. J. 1976. Breeding biology of the Grey- starlings delay incubation to avoid being cuck- facedPetrel Pterodromamacroptera gouldi. Ibis 118: olded. Auk 98: 386-389. 51-64. RICHDALE,L. E. 1963. Biology of the Sooty Shear- LUMPKINßS. 1981. Avoidance of cuckoldry in birds: waterPuffinus griseus. Proc. Zool. Soc.London 141: the role of the female. Anim. Behav. 29: 303-304. 1-117. 1983. Female manipulation of male avoid- ROBERTS,T. A., & J. J. KENNELLY. 1980. Variation in anceof cuckoldrybehavior in the Ring Dove. Pp. promiscuityamong Red-winged Blackbirds. Wil- 91-112 in Social behavior of female vertebrates son Bull. 92: 110-112. (S. K. Wasset, Ed.). New York, Academic Press. TICKELL,W. L. N. 1962. The Dove Prion, Pachyptila MACDONALD,M.g. 1975. The biology of the Fulmar desolataGmelin. Falkland Is. Dep. Surv. Sci. Rept. (Fulmarusglacialis (L)) in N.E. Scotlandwith spe- 33: 1-55. cial referenceto the pre-laying period. Ph.D. dis- 1968. The biology of the great albatrossesß sertationß Aberdeen, Scotland, Aberdeen Univ. Diomedeaexulans and Diomedeaepomophora. Pp. 1- 1977. Adult mortality and fidelity to mate 56 in Antarctic bird studies. Antarctic Res. Series, and nest-sitein a group of marked Fulmars.Bird vol. 12 (O. L. Austin, Ed.). Washingtonß D.C., Study 24: 165-168. Amer. Geophys.Union. 1980. The winter attendance of Fulmars at ß& R. PINDER. 1966. Two-egg clutchesin al- land in NE Scotland. Ornis Scandinavica 11: 23- batrosses. Ibis 108: 126-129. 29. TOMKINS,R.J. 1983. Fertilization of Wandering Al- MCKINNEY, F., K. M. CHENG,& D. J. BRUGGERS.1984. batrosseggs on Macquarie Island. Notornis 30: Sperm competition in apparently monogamous 244-246. birds. Pp. 523-545 in Sperm competition and the TRIVERS,R.L. 1972. Parental investment and sexual evolution of animal mating systems(R. L. Smithß selection.Pp. 136-179 in Sexualselection and the Ed.). New York, Academic Press. descentof man (B. Campbell,Ed.). ChicagoßA1- --, S. R. DERRICKSON,& P. MINEAU. 1983. Forced dine Publ. Co. copulationin waterfowl. Behaviour86: 250-294. WADE,M. J., & S. J. ARNOLD. 1980. The intensity of MINEAU,P., & F. COOKE. 1979. Rape in the Lesser sexual selection in relation to male sexual be- Snow Goose. Behaviour 70: 280-291. havior, female choiceßand sperm precedence. MONNETT, C., L. M. ROTTERMAN, C. WORLEIN, & K. Anim. Behav. 28: 446-461. HALUPKA. 1984. Copulation patterns of Red- WARHAM,J. 1964. Breedingbehaviour in Procellar- winged Blackbirds (Agelaiusphoeniceus). Amer. iiformes. Pp. 389-394 in Biologie Antarctique Natur. 124: 757-764. Symp. 1 (R. Carrick, M. Holgate, and J. Prevost, OGLESBY,J. M., D. L. LANIER, & D. A. DEWSBURY.1981. Eds.). ParisßHermann. The role of prolonged copulatory behavior in WARREN, D.C., & L. KILPATRICK. 1929. Fertilization facilitating reproductive successin male Syrian in the domesticfowl. Poultry Sci. 8: 237-256. golden hamsters(Mesocricetus auratus) in a com- WERSCHKUL,D.F. 1982a. Parental investment: influ- petitive mating situation. Behav.Ecol. Sociobiol. enceof nest-guardingby male Little Blue Herons 8: 47-54. Florida caerulea. Ibis 124: 343-346. OLLASON,J. C., & G. M. DUNNET. 1978. Age, expe- 1982b. Nesting ecology of the Little Blue rience and other factors affecting the breeding Heron: promiscuousbehavior. Condor 84: 381- successof the Fulmar, Fulmarusglacialis, in Ork- 384. ney. J. Anim. Ecol. 47: 961-976. WILEY,R. H. 1973. Territoriality and non-random ORING,L.W. 1982. Avian mating systems.Pp. 1-92 mating in Sage Grouse, Centrocercusurophasianus. in Avian biologyßvol. 6 (D. S. Farner, J. R. Kingß Anim. Behav. Monogr. 6: 87-169. and K. C. Parkes, Eds.). New York, Academic Press. WILLIAMS, G.C. 1975. Sex and evolution. Princetonß PARKER,G.A. 1970. Sperm competitionand its evo- New JerseyßPrinceton Univ. Press.