THE
EFFECTS
OF EXPERIENCE
PEFORMANCE
AND
AGE ON
OF WESTERN
THE BREEDING
GULLS
PETERPYLE, LARRY B. SPEAR,WILLIAM J. $YDEMAN,
AND DAVID
G. AINLEY
PointReyesBird Observatory,
StinsonBeachCalifornia94970 USA
ABSTRACT.--We
examinedthe independenteffectsof age and breedingexperienceon reproductivesuccess
in WesternGulls(Larus
occidentalis)
3-9 yearsold.Bothfactorssignificantly
enhancedsuccess
after the other factorhad been statisticallycontrolled.When the sexeswere
analyzedseparately,previousbreedingexperiencehad a significantinfluenceon success
of
femalesbut not males,whereasamongfirst-timebreeders,age was significantlymore importantto malesthan to females.The trendswith experiencewere curvilinear,which indicates
that the advantagesof experienceafter the firstbreedingattemptdecreased.
The linear trends
with age implied that increasesin success
are realized more constantlythrough 9 yearsof
age. Breedingexperienceenhancedclutchsize of the female as well as hatchingsuccess
of
bothsexes,
while ageenhancedthe timingof breedingby bothsexesandclutchsize,hatching
success,
and fiedgingsuccess
of the malebut not the female.Our resultsrelateto sex-specific
rolesin reproduction,and we suggestthat deferredbreedingmaybe moreeffectivestrategy
for malesthan for females.Received
29 November
1989,accepted
7 July1990.
REPRODUCTIVE
SUCCESS
increaseswith both age
and breeding experiencein many speciesof
birds (see summariesin Ryder 1981, CluttonBrock 1988). Becauseof small sample sizes of
marked individuals of both known-age and
known-experience, however, few studies have
consideredthe independent effectsof thesetwo
factors (see discussions in Davis 1976, Ollason
3-9 yearsold by determining eacheffectwhile
statistically controlling the other. We investigated these effects in females and males, and
we relate differencesto sex-specificroles in reproduction.Our resultsprovide somenew perspectiveson variation in reproductive performance (Nol and Smith 1987, Wooler et al. 1990)
and deferred breeding (Williams 1966).
and Dunnet 1978, Ainley et al. 1983, Saether
1990). In several species,reproductive performanceof first-time breedersincreaseswith age
(Richdale1949;Hornberger in Lack 1966;Davis
1976;Ollasonand Dunnet 1978, 1988;Harvey
et al. 1988; Wooler et al. 1990), which has
prompted speculation that age rather than
breeding experienceaccountsfor increasedreproductivesuccess
(Lack 1968,Ryder 1981;see
alsoHamann and Cooke 1987).Few studieshave
examined the contribution of prior breeding
experiencewith statisticalcontrolsfor age.
Western Gulls (Larus occidentalis)that breed
METHODS
Studyareaandfield methods.--Ainley
and Boekelheide (1990)describedthe breedingecologyof seabirds on SEFI, located 48 km west of San Francisco
and within the Californiaeasternboundarycurrent
system.Reproductivesuccess
in gulls and other seabirds is stronglyrelatedto food availability,which
in turn showswideinterannualfluctuation.
Thepopulationof WesternGullsbreedingon SEFI(est.26,000
birds)hasbeen describedby Spearet al. (1987)and
Penniman et al. (1990). As the result of an annual
chick-banding
programinitiatedin 1971,approxi-
on SoutheastFarallon Island (SEFI), California,
improve reproductive successwith age (see
Sydeman et al. 1991). In this same population
a wide variation in age of first-breedingexists,
4-10 (œ= 6.0) yr in femalesand 3-9 (œ= 4.8)
yr in males (Spear 1988, unpubl. data). Thus,
birds with varying amountsof breeding experience occur within each age group. Consequently, we examined the relative effects of
breeding experienceand age in Western Gulls
25
mately11%of the breedingpopulationwas marked
by 1983,whenwe beganour studiesof known-age
birds.Bandswere readat the onsetof territoryoccupation. Clutch initiation date, clutch size, and
hatchingand fledgingsuccess
were determinedeach
yearfora sampleof markedindividuals(seeSydeman
et al. 1991 for details). Sex of most adults was deter-
minedby behavioralobservationduring courtship
and by directsizecomparison
(Pierotti1981,Spear
1988).
Between 1983 and 1988, we monitored 3,069 breed-
The Auk 108:25-33. January1991
26
PYLEETAL.
[Auk,Vol. 108
ing attempts(definedby the laying of at leastone
egg)by 1,233known-ageindividuals(672males,521
relative contributionsof age and experiencewithin
the entire sample, we also used a hierarchical log-
females, 40 unknown sex). Within this sample, 826
linear analysis. Partial G-values were calculated for
attemptswere made by 476 individualsaged 3-9 yr
and of known breeding experience(hereafter "experience").In theseindividualsthe breedinghistory,
includingyearof firstbreeding,wasknown.We used
three different methods to determine year of first
breeding.First,454of the attemptswereby gullsthat
Spearhad markedaschicksin 1978-1980,which were
then specificallylocated through intensive searches
of the entire colonyduring the 1982-1988breeding
seasons.Second,82 attemptsinvolved 3-yr-old males
and4-yr-oldfemalesthatwere almostcertainlybreeding for the first time. In observedattemptsof >2,000
known-agegullson SEFI,no 2-yr-oldsand only one
3-yr-old female was involved (Point ReyesBird Observatory[PRBO]unpubl. data; males breed earlier
than femalesin this populationbecauseof a skewed
sexratioin breedingbirds[Spearet al. 1987]).Finally,
378attemptsresultedfrom an intensiveband-reading
effort during the 1986-1988seasons.Band numbers
interactionsbetween reproductivesuccess,age, experience,and sex.
Residuals were found to be distributed normally
for the independentvariablesin all comparisonsand
also for the dependentvariablesin analysesof reproductivesuccess
and timing of breeding(skewwas
lessthan 0.02;kurtosiswas greaterthan 2.0 and less
than 4.0 where kurtosisof normal equaled 3.0). Residuals were not distributed normally for comparisonsof clutchsize but becausethoseof the independentvariableswerenormallydistributed,no violations
of the assumptionsof linear regressionoccurred(Seber 1977: 150). Proportionaldata (hatching success
and fledging success)were also arcsine square-root
transformedbefore further analyses.
To testfor curvilinearity we performedpolynomial
regressionof trendswith respectto age and experience, to see whether or not quadraticequationssignificantly improved the fit over linear equations.
Analysesof covariancewere used to evaluate experience-specific
differencesin reproductivesuccess
with
and nest locationswere determined for every marked
bird that bred within a defined study area (g = 1,164
marked birds per year). Gulls that nestedat least 10
m within area boundariesduring 1987 or 1988 and
age,and sex-specific
differenceswithin all comparisons.
Our sample consistedof age groups 3-9 yr and
experiencegroups 0-4 yr. The effectsof age were
betweenyearsof up to severalmetersarenot unusual, testedwithin experiencegroups0, 1, and 2 yr. For
but relocations>10 m are rare (Spearet al. 1987).We analyseson the effectsof experience,we pooled age
confirmedthis by markingnestbowlsand measuring groups6, 7, and 8 yr. Although significantincreases
relocation distances of 156 banded individuals
bein reproductivesuccess
occurredbetweenagegroups
tween the 1987-1988and 1988-1989seasons.Only 3 3 and 6 in our sample,by age 6 this trend had equalhad relocated> 10m, and only one had relocated> 14 ized. No significanteffectsof agewere found between
m (mean [+SE]: 1.5 + 0.23 m; range:0.0-21.1 m). In ages6, 7, and 8 for comparisons
of reproductivesuc= 0.46, P = 0.500) or
all casesit wasalsoknown that birds bred during all cess(X• = 0.77, P = 0.943; F•,3o?
interimyears.Threeindividualsof oursampleskipped of any of its four components(X2 < 6.3, P > 0.3; F <
a year(failed to lay eggs)betweenattempts,although 2.7, P > 0.1).
To ensurethat our samplewas distributedproporall threemaintainedterritoriesthroughouttheskipped
season.As it cannotbe judgedhow much experience tionally among years of varying food supply (see
isgainedduringskippedyears,thesethreebirdswere Sydemanet al. 1991),we usedanalysesof covariance
to control for year, and we obtained resultssimilar
subsequentlyexcludedfrom the sample.
Dataanalysis.--Wedefinedreproductive
success
asthe to thosepresentedfor all examinationsof both age
number of chicksfledgedper breeding attempt.Four and experience. We also examined the effects of excomponentsof reproductivesuccesswere also ex- periencewithin a targetedsubsample(135 attempts)
amined in sex-specificanalyses:timing of breeding of 6-yr-oldsbreeding in 1988,which controlledfor
(indicated by clutch initiation date), clutch size, both year and cohort.
In the sampleof gulls aged 6-8 yr that were used
hatchingsuccess
(%of eggslaid thathatch),andfledgin the experienceanalyses,24% of the attemptsining success
(% of chickshatchedthat fledge).
Linear regressionson age- and experience-con- volved the sameindividual in different years.To introlled sampleswere usedfor mostcomparisons.Sig- vestigatethe effect of dependenceamong observanificance is indicated by P values on the F-statistic. tions, we used a random numbers table to eliminate
We performed ANOVA and Chi-square analyses, multiple attemptsby the same individual. We again
which yielded similar resultsfor all comparisons.
We found similar levelsof significancebetweenanalyses
presentresultsof regression
analysisbecauseit is a on thefull (includingmultipleattempts)andreduced
that were not recorded the previous year were as-
sumed to be first-time
breeders. Nest-site
movements
morepowerfulanalysis
whentheindependent
variables(ageandexperience)
arequantitative(Breslow
andDay1980,GainesandRice1990).To confirmthe
resultsof the regression
analyses
and to examinethe
(one attempt per individual) samples;data from the
largersamplesarepresented.All statistical
testswere
performedusingthe SPSSPC
+ program;significance
is implied when P < 0.05.
January
1991]
Experience
andAgein Western
Gulls
27
2.5
..& (4)
o
2.0
(46)
LU
o
o
o-
""
(46) ./'
•r ./-
19
j
(37)
1.0
(791
ß overall group; P=O.002
0.5
ß
o_
LLI
6-yr-olds
in 1988; P=O.001
0.0
0
1
YEARS
2
BREEDING
3
4
EXPERIENCE
Fig. 1. The effectsof breedingexperienceon reproductivesuccessin 6-8 yr olds (overall) and 6-yr-olds
breedingin 1988(œ+ I SE).Reproductivesuccess
is definedas chicksfledgedper breedingattempt.Sample
sizesare in parentheses.
Significanceis determinedwith linearregression.
Seetext for resultsof analysesof
curvilinearity.
although this trend was not curvilinear (P =
0.08), perhapsbecauseof smallersamplesizes.
Breeding
experience
andreproductive
success;
age The increasein success
with experienceis prescontrolled.--Amongindividuals aged 6-8 yr, re- ent when cohort and year, as well as age, are
productivesuccess
increasedsignificantlywith controlled.
experience (Fig. 1). Most of the increase ocThe positive influenceof experienceproved
curred between the first and secondbreeding significantwith femalesbut not with males(Fig.
attempt. This resulted in a significantcurvilin- 2), although analysisof covarianceshowedno
ear trend (P = 0.002). A significantcorrelation difference between the sexes(F•,4•5= 2.49, P =
alsoexistedin the 1988subsampleof 6-yr-olds, 0.116). In females, reproductive successinRESULTS
•
(16)
2.0
(49)
LU
o
(74)
7
(19)
1.5
(40)
'
(69
uJ
-> 1.0
I--
/
(34)
(39)
o
0.5
ß females; P=O.004
ß males; P=0.146
i-i
LU
0.0
0
1
YEARS
BREEDING
2
3
4
EXPERIENCE
Fig. 2. The effectsof breeding experienceon reproductivesuccessof malesand females6-8 yr old (œ+
1 SE).Reproductivesuccess
is defined aschicksfledgedper breedingattempt.Samplesizesare in parentheses.
Significanceis determinedwith linear regression.Seetext for resultsof analysesof covarianceand curvilinearity.
28
PYLE
ETAL.
[Auk,Vol. 108
TABLE1. Log-linear analyses on number of chicks sexesgrouped and separated,and it confirmed
fledged. Interactions of reproductive successwith that the relationshipof success
with experience
age and experienceare given for the entire popudid
not
significantly
vary
with
sex (Table 1).
lation and separatelyfor eachsex.Interactionswith
Timing of breedingand fledgingsuccess
were
not significantly related to previous breeding
covariance. Twenty-seven cases (16 male, 11 fe- experiencein either sex, although a tendency
male) were rejectedbecauseof missingdata points. for later breeding and poorer fledging success
by less-experienced
femaleswas evident (Table
df
Partial G
P
2). Clutch size increasedsignificantly with exAll birds (799)
perience in females but not in males, and the
Age
10
47.984
0.001
sex-specificdifferencein this trend was signifExperience
6
37.413
0.001
icant. As with overall reproductivesuccess,
the
Age by sex
6
76.169
0.001
largest increase in clutch size of females ocExperienceby sex
6
16.544
0.198
curred between the first and secondbreeding
Females (297)
attempts. Hatching successincreased signifiAge
9
15.984
0.594
cantly with breeding experienceof both sexes,
Experience
6
27.746
0.006
with
the largest increasesagain occurring beMales (502)
tween the first and secondattempts.Thus, the
Age
10
44.329
0.001
overall increasein reproductive successresultExperience
6
18.918
0.090
ed primarily from increasedclutch size of experiencedfemalesand improved hatching succreased through 4 yr of experience, but es- cesswith experienceof both sexes.
peciallybetweenthe first and secondbreeding
Ageandreproductive
success;
breeding
experience
attempts.This trend was significantly curvilin- controlled.--Wetested the effectsof age on reear (P = 0.008). In males successactually de- productivesuccess
within experiencegroupsof
clined after two yearsof experience,although 0, 1, and 2 yr (sexesgrouped; Fig. 3). Age corthis declinewasnot significant(P = 0.146).The related significantly with reproductivesuccess
log-linear analysison the entire sample con- during the first breeding attempt;most of the
firmed the resultsof linear regressionwith the difference occurredbetween the agesof 3 and
age and experienceby sex are also given for the
entire sample,to confirm the resultsof analysesof
TABLE2. The effectsof previousbreeding experienceon componentsof reproductivesuccessof malesand
females (œ+ SE). Significancewas determined with linear regression(P•). P2 comparesthe regressionof
males and femalesusing analysisof covariance.Experiencegroups 3-4 are lumped in the table although
regressions
were performedwithout pooling.Proportionaldata(hatchingand fledgingsuccess)
were arcsine
square-roottransformed before analyseswere performed. Sample sizes (in parentheses)for each mean are
+ 2 becauseof missingdata points.
Years Breeding Experience
0
I
2
3-4
(39•, 40•)
(48•, 69•)
(44•, 74•)
(35•, 72•)
•
P•
133.1 + 1.3
132.3 + 1.3
132.0 + 0.9
130.0 + 0.7
130.1 + 0.8
129.6 + 0.6
130.2 + 1.0
130.1 + 0.8
-0.93
-0.01
0.069
0.986
0.157
P2
Timing a
Females
Males
Clutch
size
Females
2.56 + 0.10
Males
2.75 + 0.08
2.83 + 0.06
2.91 + 0.05
2.89 _+ 0.05
2.77 + 0.06
2.89 _+ 0.06
2.83 + 0.05
0.11
-0.01
0.004
0.724
0.011
0.885 _+ 0.03
0.797 _+ 0.04
0.837 + 0.04
0.836 _+ 0.03
0.919 + 0.04
0.890 _+ 0.03
0.12
0.08
0.001
0.004
0.415
0.653 _+ 0.06
0.672 _+ 0.04
0.724 _+ 0.06
0.721 _+ 0.04
0.677 _+ 0.06
0.647 + 0.05
0.04
0.00
0.220
0.923
0.354
Hatching success
Females
Males
0.598 _+ 0.06
0.738 _+ 0.06
Fledging success
Females
Males
0.575
_+ 0.07
0.581 _+ 0.07
Timing is representedas the Julian date of clutch initiation.
January1991]
Experience
andAgein Western
Gulls
ß exp=O;
0'•
LLI
2.0
0'•
29
P=O.001
ß exp=l; P=0.104
(60)
ß exp=2; P=0.482
W
(82)
(22)
I ..-'(1
I--
•
lO)
r,.)
O
n"
(61)
•11 (71)
0.5
(64),
n- 0.0 ------•
5
I
I-
I
I
4
5
6
7+
AGE
(YEARS)
]Fiõ.3. The effectsof aõe on reprociucfiYe
success
in three experienceõroups(f _+1 SE).ReprociuctiYe
success
is ciefineci
as chicksfleciõeci
per breeciinõattempt.Samplesizesare in parentheses.
Siõnificance
is
cietermineci
with linear reõression.
Seetext for resultsof analysesof coYariance
anti curvilinearit¾.
4 yr. The effect of age on reproductivesuccess ence,age and reproductivesuccessrelated sigwas not significant in the 1-yr and 2-yr ex- nificantly in malesbut not femalesduring the
perience groups (Fig. 3), although analysisof first breedingattempt (Fig. 4). This sex-specific
covariance showed no differences between the
differencewassignificantbothoverall (analysis
trends of the three groups (F2,665
= 0.43; P = of covariance,
F•,3•5
= 10.15,P = 0.002),and when
0.651).The trend in first-timebreederswassig- 3-yr-old maleswere excluded (F•,•s•= 4.52; P =
nificantly curvilinear (P = 0.035). We suspect 0.034).The latter analysisensuredthat the overthat this resulted from the combination
of the
all effectwas not solelydue to the factthat males
separate,noncurvilinear trends observed when
the sexeswere separated(seeFig. 4).
In contrastto the effectsof breeding experi-
but not femalesbreedat age 3. In males,most
of the increaseoccurredbetween3- and 5-yrolds, but interestinglythis trend was not sig-
2.0
ß females;
w
1.5
P=0.482
(4)
ß males;
P=O.001
(25)
(40)
•r•
(36)
I
ß
LU 1.0I-
0
,.-'
0.5
,," 1
(25)
rr
n
W
rr
0.0
5
4
5
AGE
6
7+
(YEARS)
Fig. 4. The effectsof age on reproductivesuccess
in first-timebreedingfemalesand males(œ+ 1 SE).
Reproductive
success
is definedas chicksfledgedper breedingattempt.Samplesizesare in parentheses.
Significance
is determinedwith linearregression.
Seetextfor resultsof analyses
of covariance
and curvilinearity.
30
PYLE
ETAt,.
[Auk,Vol.108
TABLE
3. The effectsof ageduringthe firstbreedingattempton components
of reproductivesuccess
of males
and females(•?+ SE).Significancewasdeterminedwith linear regression(P•).P2comparesthe regressions
of malesand femalesusing analysisof covariance.Age groups 6-9 are lumped in the table, although
regressions
were performedwithout pooling.Proportionaldata(hatchingand fledgingsuccess)
were arcsine
square-roottransformedbefore analyseswere performed.Samplesizes(in parentheses)for each mean are
+4 becauseof missingdata points.
Age (yr)
3
4
(568)
(39•, 688)
5
6-9
(42•, 2•)
(39•, 408)
•
P•
P2
-1.92
-3.48
0.009
0.001
0.127
Timing a
Females
Males
Clutch
-142.9 + 1.3
137.3 + 1.2
134.9 + I.I
133.7 + 0.1
133.1 + 1.5
133.1 + 1.3
132.3 + 1.3
size
Females
Males
-2.04 + 0.08
2.44 + 0.10
2.44 + 0.07
2.57 + 0.09
2.54 + 0.13
2.56 + 0.10
2.75 + 0.08
0.06
0.21
0.257
0.001
0.033
0.645 + 0.06
0.615 + 0.05
0.746 + 0.05
0.750 + 0.06
0.598 + 0.06
0.738 + 0.06
0.04
0.07
0.306
0.011
0.467
0.583 + 0.08
0.490 + 0.06
0.447 + 0.06
0.561 + 0.09
0.575 + 0.07
0.581 + 0.07
0.04
0.09
0.348
0.001
0.039
Hatching success
Females
Males
-0.467 + 0.06
Fledging success
Females
Males
-0.278 + 0.07
Timing is representedas the Julian date of clutch initiation.
nificantly curvilinear (P = 0.084). The sex-specificdifferencesand ageeffectson reproductive
successwere confirmed by the log-linear analysis using the entire sample (Table 1).
All four componentsof reproductivesuccess
contributedto the age-relatedincreasefound
in males, but only timing of breeding related
significantly to age in first-breeding females
(Table 3). The differences between the sexesin
trendsof clutchsize and fledging success
were
significant.Hence the increasein reproductive
successwith age of first-time breedersresults
primarily from increasedclutch size, hatching
success,and fledging successwith age of the
male. Earlier breeding by older individuals of
both sexesmay alsocontribute to increasedsuc-
ern Gannets(Morusbassanus)
probably resulted
from a maturationof innate qualitieswith age
as opposedto a learning processthrough increasedbreedingexperience.Wooleret al. (1990)
concludedthat age rather than breeding experience enhanced reproductive successin
Short-tailedShearwaters(Puffinus
tenuirostris).
It
was suggested that "environmental" experience, or age, rather than breeding experience
accountedfor increasesin clutchsize with age
of female Ad•lie Penguins(Pygoscelis
adeliae;
Ainley et al. 1983), Snow Geese(Chencaerulescens; Hamann and Cooke 1987), and Brandt's
Cormorants (Phalacrocorax
penicillatus;Boekelheide and Ainley 1989). In Western Gulls, on
the other hand, we have shown that both age
and breedingexperienceseparatelyenhancereproductive success.Furthermore we found that
DISCUSSION
clutchsize increasedwith breedingexperience
rather than ageof femaleWesternGulls,which
Ageversus
experience.--On
the basisof studies suggeststhat in this speciesreproductive exthat find improvement with age in the success periencehasa greaterinfluencethan ageon the
of first-time breeders, Lack (1968) and Ryder reproductive condition of the female (see Ain(1981) speculatedthat previous breeding ex- ley et al. 1983).We believethat hatchingsuccess
periencedoesnot enhancereproductiveperfor- is influencedby learning through experience,
mance. Rather, observedincreasesin reproduc- while timing and fledging success
are more aftive successare strictly age-related or at least fected by age. The curvilinear relationshipsof
related to experience not associatedwith nest- success
with experiencefurther imply that the
ing. Nelson (1966) suggestedfurther that in- advantagesof this learning processbecome
creasedreproductivesuccess
with agein North- smaller with each additional year after first
cess.
January
1991]
Experience
andAgein Western
Gulls
breeding, whereas the linear trends with age
indicate that age-related factorscontinue to influence success,at least through 9 yr of age.
Breedingfailure of individual WesternGulls
hasbeen ascribedlargely to egg neglect,infertility, and intraspecific predation on eggs and
young chicks (Coulter 1973, Spear et al. 1987,
Penniman et al. 1990). These components of
reproductive successmay, in general, be reducedby increasedexperience,whereastiming
and foraging efficiencyseemto improve more
with age. In other species,such as those mentioned above, reproductive successcould be
moredependenton timing and food acquisition
than on egg loss and predation, which might
accountfor the implicit higher correlationwith
age than with experience in these species.Unlike the penguinsand cormorants,WesternGulls
on SEFI also have a high rate of mate fidelity
(PRBO unpubl. data), which may amplify the
effect of previous breeding experience (see
Coulson 1966). In situations where adult sur-
vivorshipand matefidelity are lower, breeding
experiencemay have lesseffecton reproductive
Sl•CCeSS.
Females
versus males.--We
found
that
in-
creasedbreeding experience significantly enhancesthe reproductivesuccessof femalesbut
not males,whereasage is significantlymore important to males. Other studies (Davis 1976,
Thomasand Coulson1988,Sydemanet al. 1991)
that did not differentiate between age and experiencefound no differencesbetweenfemales
and maleswhen correlatingeither of thesefactors with reproductivesuccess.In our study, it
was only after age and experiencewere alternately controlled that the difference between
the sexesbecameapparent.
We believe that this sex-specificeffectresults
primarily from differencesin parental rolesin
reproduction.Pierotti (1981) found that female
Western Gulls on SEFI typically spend more
time incubatingand directlyprotectingthe eggs
and small chicks,while malesspend more time
in territorial defenseand feed the young larger
amounts
of food than
the females.
It follows
that hatching success,which is strongly influ-
encedby nest-sitebreedingskills(Nelson 1966,
Ollason and Dunnet 1986), increasessignificantly with experienceof the female and to a
lesser degree the male, while the increasein
fledging successobserved for older males resuitsin part from an increasein foragingskills
with age (Nur 1984;seealso Searcy1978, Greig
31
et al. 1983, MacLean 1986). The significant increasein clutch size with age of male Western
Gulls may alsoresult from the amount of food
fed to the female during courtship.Again this
relatesto increasedforaging efficiency.In contrast to the female's nest-site skills, the male's
foraging experienceincreasescontinually with
age, but would not be influenced as much by
increasedbreeding experience.
Life-historystrategy.--Four hypotheseshave
been proposed recently to explain increasesin
reproductivesuccess
with age in birds (Nol and
Smith 1987, Wooler et al. 1990). In addition to
higher successdue to increasesin reproductive
efficiencywith (1) breedingexperience
andwith
(2) chronological age per se, it has also been
suggestedthat (3) older birds increasereproductive effort to offset decreased survival (re-
sidual reproductive value) with age, and that
(4) apparent increaseswith age are simply an
artifact of increased survival of higher-quality
individuals, which have higher rates of reproductive performanceirrespectiveof age or experience. Our results are consistent with the
first two hypotheses.Although in the present
analysiswe lack the data on older birds needed
to fully test the third hypothesis,the experience-, age-, and sex-relatedvariation in the effects of the four componentsof reproductive
successimplies that the increasein reproductive successin Western Gulls resultsprimarily
from a combinationof increasein foragingskills
with age, and improvement of other nesting
skills through increased breeding experience
(Nur 1984, Reid 1988), rather than through a
general increasein reproductiveeffort (Pugasek
1981, 1984). As for the fourth hypothesis,the
curvilinear effect of breeding experiencewhen
age was controlled suggeststhat the effect of
breeding experience is real rather than an artifact of differential
survival
rates. However,
a
costof reproduction(if it existsin this species)
would complicatethis relationship.The linear
increasein success
with age in first-time breeders and the sex-specificdifferencesalso imply
that the effectof agein malesis real.We suspect
that quality of individual and differentialsurvival may partially accountfor observedincreasesafter the first breeding attempt, perhaps
more influentially with females.We are currently collectingthe datato examineolder age
groupsof WesternGulls,and in a separateanalysis, we will considerthe effectsof age and
experienceon survivorship.
32
PYLE
ETAL.
Deferredbreeding.--Toexplain deferred maturation and low-intensity breeding in seabirds,
Williams (1966; see also Nur 1988) suggested
that the costof reproductiveeffort jeopardizes
future reproductiveoutput by increasingthe
parental risk of mortality. Other researchers
(Lack1968,Goodman1974,Chabrzykand Coulson 1976,Ainley et al. 1983,Hamann and Cooke
1987,Ainley et al. 1990)havefurther suggested
that breeding at a young age is disadvantageous, as the costsin terms of reduced survivorshipdo not justify the low reproductiveoutput of younger birds. Assumingthesestrategies
are adaptive,we suggestthat deferredbreeding
in Western Gulls is more advantageousto males
than to females.By gaining more foragingexperiencebefore attempting to breed, malesenhance their chances of reproductive success
without incurring the costsof potentially unsuccessfulbreeding.
For females,by contrast,initial low-intensity
breeding may be more adaptative, as females
more than males gain from the experienceof
earlier reproductiveattempts.In contrastto foraging skillsßcompetencein incubatingeggsand
caring for newly hatched chicks is gained
through breeding experience,regardlessof age
(see Curio 1983). The improvements in reproductive success
gained through deferredbreeding in females were negligible. The fact that
first-breedingfemalesbut not maleshad smaller clutches
indicates
that
initial
costs are min-
imized by females through a low-intensity effort.
[Auk,Vol.108
LITERATURE CITED
AINLEY,D. G., R. E. LEREOCHE,
& W. J.L. SLADEN.1983.
Breedingbiologyof the Ad•lie Penguin.Berkeley, Univ. California Press.
, & R. J. BOE•'•HœIDœ(Er•s.). 1990. Seabirds of
the Farallon Islands:ecology,structureand dynamics in an upwelling system. Palo Alto, Stanford Univ.
Press.
ß C. A. RIBIC, & R. C. WooD.
1990. A demo-
graphic study of the South Polar Skua Catharacta
maccormicki
at Cape Crozier. J. Animal Ecol. 59:
1-20.
BOEKELHEIDE,
R. J., & D. G. AINLEY. 1989. Breeding
biology of the Brandt's Cormorant: interaction
between age and resourceavailability. Auk 106:
389-401.
BRESLOW,N. E., & N. E. DAY. 1980. Statistical meth-
odsin cancerresearch.
Vol. 1,theanalysisof casecontrol studies.Lyon, France, Int. Agency Res.
Cancer.
CHABRZYK,G., & J. C. COULSON. 1976. Survival and
recruitmentin the Herring Gull Larusargentatus.
J. Animal Ecol. 45: 187-203.
CLUTTON-BROCK,
T. H. (El:).). 1988. Reproductivesuccess: studies of individual
variation
in contrast-
ing breeding systems.Chicago, Univ. Chicago
Press.
COULSON,
J.C. 1966. The influenceof pair bondand
ageon the breedingbiology of the kittiwake gullß
Rissatridactyla.J. Animal Ecol. 35: 269-279.
COULTER,
M. C. 1973. The breeding biology of the
Western Gull (Larusoccidentalis).
M. S. thesis,Oxford, Oxford Univ.
CURIO,E. 1983. Why do young birds reproduceless
well? Ibis 125: 400-404.
DAvis,J. W. F. 1976. Breedingsuccess
and experiencein the ArcticSkua,Stercorarius
parasiticus
(L.).
J. Animal Ecol. 45: 531-536.
GAINES,
S. D., & W. R. RICE. 1990. Analysisof biologicaldatawhen there are orderedexpectations.
ACKNOWLEDGMENTS
Our work on SEFI has been enthusiasticallysupported by the Farallon Patrol of Ocean Alliance, the
U.S. CoastGuard, and especiallythe U.S. Fish and
Wildlife Service, which administers the Farallon Na-
tional Wildlife Refuge. Former and present Farallon
biologistswere instrumentalin initiating and devel-
oping the WesternGull project;we especiallythank
R. Boekelheide, J. Penniman, T. Penniman, H. Carter,
and S. Eroslie. Much of the fieldwork
was carried out
by Farallon volunteers;we would like to especially
acknowledge the efforts of C. Askew, L. de Forrest,
K. Donaghue, S. Johnston,K. Lee, M. Lerdau, J. Lyons, M. Rossi, and C. Valle. Significant help with
statisticswas offered by N. Nut and L. Stenzel, and
the comments of S. Eroslie, N. Nur, R. Pierotti, and
an anonymous reviewer improved the manuscript.
This is Contribution No. 448 of the Point ReyesBird
Observatory.
Am. Nat.
135: 310-317.
GOODMAN, D. 1974. Natural selection and a cost ceil-
ing on reproductive effort. Am. Nat. 108: 247268.
GREIG,S. A., J. C. COrn,SON,& P. MONAGHAN. 1983.
Age-relateddifferencesin foragingsuccess
in the
Herring Gull. Animal Behav.31: 1237-1243.
H•N,
J., & F. COOKE.1987. Age effectson clutch
sizeand laying datesof individual female Lesser
Snow Geese Anser caerulescens. Ibis 129: 527-532.
HARVEY,P. H., M. J. STENNING,& B. CAMPBELL.1988.
Factorsinfluencing reproductivesuccess
in Pied
Flycatcher.Pp. 189-200 in Reproductivesuccess:
studies of individual variation in contrasting
breeding systems(T. H. Clutton-Brock,Ed.). Chicago,Univ. ChicagoPress.
LACK,D. 1966. Population studiesof birds. Oxford,
Clarendon
Press.
January
1991]
Experience
andAgein Western
Gulls
1968. Ecologicaladaptionsfor breeding in
birds. London, Methuen
& Co.
33
RICHDALE,
L.E. 1949. A studyof a groupof penguins
of known-age. Dunedin, New Zealand, Biol.
MAcLF.
AN,A. A.E. 1986. Age-specificforagingability and the evolution of deferred breeding in
three speciesof birds. Wilson Bull. 98: 267-279.
NELSON,
J.B. 1966. The breedingbiologyof the Gan-
Monogr. No. 1.
RYDER,
J.P. 1981. The influenceof age on the breeding biologyof colonialnestingseabirds.Pp. 153-
net Sula bassanaon the BassRock, Scotland. Ibis
birds(J.Burger,B.L. Oilan, and H. E.Winn, Eds.).
108: 584-626.
New York; Plenum Publ. Co.
NOL, E., & J. N.M. SMrrH. 1987. Effectsof age and
breeding experienceon seasonalreproductive
successin the Song Sparrow. J. Animal Ecol. 56:
301-313.
S.•'rH•R, B. 1990. Age-specific variation in repro-
ductiveperformanceof birds.Pp.251-283inCurrent Ornithology, vol. 7 (D. Power, Ed.). New
York, Plenum Publ. Co.
NUR, N. 1984. Increasedreproductivesuccesswith
age in the California Gull: due to increasedeffort
or improvement of skill? Oikos 43: 407-408.
1988. The costof reproductionin birds: an
examination
of the evidence. Ardea 76: 155-168.
OLLASON,J. C., & G. M. DUNNET. 1978. Age, experience, and other factorsaffecting the breeding
successof the Fulmar, Fulmarusglacialis,in Orkney. J. Animal Ecol. 47: 961-976.
--,
&
. 1986. Relative effectsof parental
performanceand eggqualityon breedingsuccess
of FulmarsFulmarusglacialis.Ibis 128:290-296.
, & --.
1988. Variation in breeding successin Fulmars.Pp. 268-278 in Reproductivesuccess: studies of individual
variation
in contrast-
ing breedingsystems(T. H. Clutton-Brock,Ed.).
Chicago, Univ. Chicago Press.
BOEKELHEIDE.
1990. Western Gull. Pp. 218-244
inSeabirdsof the FarallonIslands:ecology,structure and dynamicsof an upwelling systemcommunity (D. G. Ainley and R. J.Boekelheide,Eds.).
Palo Alto, Stanford Univ. Press.
PIEROT'n,
R. 1981. Male and female parental rolesin
the Western
Gull
under
different
SEARCY,
W. A. 1978. Foraging successin three age
classesof Glaucous-wingedGulls. Auk 95: 586588.
SEBEI•,
G. A.F. 1977. Linear regressionanalysis.New
York, John Wiley & Sons.
S?F.
AR,L.B. 1988. Dispersalpatternsof WesternGulls
from
environment
Auk 98: 532-549.
Science 212: 822-823.
1984. Age-specificreproductivetacticsin the
California
Farallon
Island.
Auk
105: 128-
, H. R. CARTER,T. M. PENNIMAN,J.P. PENNIMAN,
& D. G. AINLEY. 1987. Survivorship and mortality factorsin a population of Western Gulls.
Pp. 44-56 in Studies in Avian Biology, vol. 10
(J. L. Hand, W. E. Southern, & K. Vetmeet, Eds.).
Lawrence, Kansas, Allen Press.
SYDEMAN,W. J., J. F. PENNIMAN, T. M. PENNIMAN, P.
PYLE,& D. G. AINLEY. 1991. Breedingperforage,timing of breeding,and year in relation to
food availability.J. Animal Ecol.In press.
THOMAS,C. S., & J. C. COULSON.1988. Reproductive
success
of kittiwake gulls,Rissatridactyla.Pp. 251262 in Reproductivesuccess:
studiesof individual
variation in contrastingbreedingsystems(T. H.
Clutton-Brock,Ed.).Chicago,Univ. ChicagoPress.
WILLIAMS, G.C.
Gull. Oikos 43: 409-410.
REID, W. V. 1988. Age-specificpatterns of reproduction in the Glaucous-wingedGull: increased
effort with age?Ecology69: 1454-1465.
1966. Natural selection, the costs of
reproduction,and a refinement of Lack'sprinciple. Am. Nat. 100: 687-690.
PUGASEK,
B. H. 1981. Increasedreproductiveeffort
with agein the California Gull (Laruscalifornicus). WOOL•,
--.
Southeast
141.
mance of the Western Gull: effects of parental
PENNIMAN,T. M., M. C. COULTER,L. B. SPF,•R,& R. J.
conditions.
164 in Behavior of marine animals. Vol. 4, marine
R. D., J. S. BRADLEY,I. J. SKIRA, & D. L.
SERVENTY.1990. Reproductive successof Shorttailed ShearwatersPuffinustenuirostris
in relation
to their age and breeding experience.]. Animal
Ecol. 59: 161-170.