RECRUITMENT
REPRODUCTION
AND
THE
TIMING
IN LESSER SNOW
OF
GEESE
(CHEN CAERULE$CEN$ CAERULE$CEN$)
FRED COOKE,• C. SCOTT FINDLAY, • AND ROBERTF. ROCKWELL2
•Departmentof Biology,Queen'sUniversity,Kingston,OntarioK7L 3N6 Canada,and
2Department
of Biology,City Collegeof New York, New York, New York 10031USA
A•3STRACT.--Recruitment
of offspringinto a breedingpopulationof LesserSnow Geese
(Chencaerulescens
caerulescens)
at La PerouseBay, Manitoba was usedas a measureof reproductive successto assessthe relative fitness of females who hatched their clutchesearly,
middle, and late in the breeding season.In three of seven seasonsinvestigated,goslings
from early-hatchingclutchesshowed significantlygreaterrecruitment ratesthan their middle- or late-hatchingcounterparts.No significantdifferencesin recruitmentrateswere detectedin the other four seasons,
althoughearly-hatchingclutchesshowednumericallyhigher
recruitment rates in three of these seasons.There is, therefore, some indication of directional
selectionfor early breeding.This conclusioncontrastswith that drawn by Cookeand Findlay
(1982),who, usingfledgingsuccess
asa measureof reproductivefitness,showedthat females
whoseclutcheshatchedin the middle period had the highestfitnessand concludedthat the
populationwas being exposedto stabilizingselectionfor synchronization.The discrepancy
between these resultsand thosepresentedin this paper indicatesthat conclusionsconcern-
ing the actionof selectionin naturalpopulationsdependheavilyupon the stageof the life
cycleduring which reproductivesuccess
is estimated.As such,evolutionarybiologistsmust
be cautiousof relying too heavily on measurestakentoo early in the life of the organism.
Received28 June1983, accepted12 January1984.
MEASURES
of the reproductive fitnessaccrued pattern hasbeen documentedin Common Oysostralegus;
Harris 1969),
to different segmentsof a population often pro- tercatchers(Haernatopus
vide insight into the actionof natural selection. Black-leggedKittiwakes (Rissatridactyla;CulEver since the writings of Darwin on sexual
selection, biologists have been interested in in-
len, in Perrins 1970), and Razorbills (Alca torda;
Lloyd 1979) among others. In other species,
fledging successand clutch size have different
breeding within a season. Darwin (1871) ar- intraseasonalpatterns.For example,Findlay and
gued that birds breeding early in the season Cooke (1982b)demonstratedlarger brood sizes
had an advantageover later breeding ones in at fledging for those females whose clutches
that more offspring of higher quality could be hatched near the population mean, i.e. during
raised. Various measuresof reproductive fit- the middle hatch period.
Clearly, mean brood size at fledging is a valness have been used to assessthe validity of
Darwin's generalization. A seasonaldecline of id measureof reproductivesuccessonly if surclutch size with hatch date has been demonvival to breeding age is similar among all segstratedfor severalavian species(seeKlomp 1970 ments of the fledgling population. This
for review), but it is by no meansthe only pat- assumptionis often made, but rarely justified.
tern. Blackbirds(Turdusrnerula;Snow 1958) and While offspring recruitment into the breeding
terindividual variability in the timing of
Robins (Erithacus rubecula; Lack 1946) have
population is the best available measure of an-
smaller clutch sizesat both the beginning and
end of the breeding season.A seasonaldecline
in reproductive fitness has also been shown
(though lessfrequently) when fiedging success
nual reproductive success,sample sizes are
usually too small to make meaningful compar-
is used as the measure
cruitment.
of fitness.
O'Donald
isons. Perrins (1965) used survival to at least 3
months of age as an alternative measureto reHis conclusions
were
based on the
(1972) and O'Donald et al. (1974), for example, percentage of banded fledglings retrapped or
showed a decline in the relative fledging suc- recoveredat least 3 months after fiedging and
cessof ParasiticJaegers(Stercorarius
parasiticus) included both individuals capturedduring the
as the breeding seasonprogressed.A similar winter and those trapped as breeding birds on
451
The Auk 101:451-458. July 1984
452
COOKE,
FINDLAY,
ANDROCKWELL
the study area. Using this measure,he showed
the greatestsurvival among the earliest hatching young. Lack (1966) using a similar measure
in his analysisof Pied Flycatcher(Ficedulahypoleuca)data collected by Campbell, showed
highest recapturerates for birds hatched early
in the season.Becausemost fledglings recaptured as adultswere breedersin the studyarea,
this measure approximates true recruitment
rates, although it must be recognized that recruits from a particular age cohort will not necessarilyall enter the breeding populationat the
same age. To our knowledge, however, pure
recruitment
has been
used to assess the influ-
ence of the timing of reproduction on fitness
only by van Noordwijk et al. (1980) in their
studyof Dutch populationsof the Great Tit (Parus major). Unfortunately, their conclusions
were somewhat limited by small sample sizes
despitehaving 16 yr of useabledata.
In this paper we investigatethe relationship
between the timing of reproduction and recruitment
rates of female
Lesser
Snow
Goose
( Chencaerulescens
caerulescens)
goslingsinto the
breeding segmentof the La PerouseBay colony
near Churchill,
Manitoba, Canada.
METHODS
The La PerouseBay Snow GooseProjecthas been
in continuous operation since 1968. It was not until
1973, however, that adequate numbers of goslings
were individually marked at hatch to allow for subsequent identification in the breeding population.
Eachseason,up to 2,500 nestsare found and visited
during the hatchingperiod (usuallymid- to lateJune).
Hatching of the eggswithin the clutchis highly synchronized (Syroechkovsky1975, Cargill and Cooke
1981).At the time of hatch,goslingsare marked with
small individually numbered metal web-tags,which
allow for subsequentidentification. From 1976 onwards, the web-tag code also indicated the year in
which the tag was applied, becauseit was realized
that the legible tags often remained in the web for
severalyears.Approximately5,000-6,000goslingsare
web-taggedeachyear. Someof theseare subsequent-
ly roundedup, captured,and bandedshortly before
fledging (about 6 weeks post hatch). These birds receive individually identifiable coloredleg bandsso
that if they return to the colony, they can be identified at a distance.
Recruitmentof goslingsinto the breeding colony
occursgenerally between 2 and 5 yr of age (Finney
and Cooke 1978). Recruits are identified in one of
three ways. They may be sightedat a nest during the
incubation or hatching period. They may be sighted
[Auk, Vol. 101
with goslingsduring the post-hatchperiod from observation towers set at the periphery of the feeding
areas.Only leg-banded birds (i.e. birds web-tagged
on the day of hatch and subsequentlycolor banded
5-6 weeks later) can be identified by thesetwo methods. The third method involves individuals captured
during the annual banding drives. Most nonbreed-
ing geese leave the breeding grounds on a northward molt migration (Abraham 1980),such that virtually all adult geese captured in these drives are
breeding birds. Becausethese individuals are examined in the hand, they can be identified as originating from the La PerouseBay colony either if they
have a web-tag or if they are carrying leg bands,
which indicate the bird had a web-tag when origi-
nally banded. The tags often remain in the web for
severalseasons,and it may be severalyearsafter hatch
before a bird is first found as a recruit. For goslings
taggedafter 1976,it is possibleto identify their natal
cohort basedupon the tag alone. Unfortunately, goslings taggedbefore 1976 must have been recaptured
and banded (in the same year as they were tagged)
in order to be identified as affiliated with a particular
natal cohort. Only goslings resighted or recaptured
at least2 yr after their original tagging date are considered to have been recruited into the breeding segment of the population.
Not all recruits that return to the breeding colony
are recognizedas such.A bird that carriesa web-tag
but no color band will not be recognized at a nest.
Similarly, a bird whose nest fails early in the nesting
period will be neither seen nor recaptured.For this
reason, we define the term recruitment as the num-
ber of recruitsdetectedfrom a particular cohort (i.e.
age class)relative to the number of goslingsof that
age classthat were web-tagged.This value is inevitably considerablysmaller than the true frequencyof
recruits.When goslingsare divided into three groups
(thosefrom early-, mid-, and late-hatchingnests),the
detected recruitment rate of each group is measured
relative to the number of web-tags applied to that
group. There is no reasonto believe that the detectability of recruits differs among early-, mid-, and
late-nesting birds.
RESULTS
In Snow Geese, strong female natal philopatry combinedwith a lack of male natal philoparry (Cooke et al. 1975) means that the vast
majority of the recruits from the natal cohorts
are female. In the following analyses we are
therefore concernedonly with female recruits.
Table
1 shows
the
detected
recruitment
rates
for each gosling cohort (1973-1979). To ensure
sufficientsample sizes,we have restrictedour
analysesto thosecohortsincluding at least 45
July1984]
Recruitment
andTiming
inSnow
Geese
TABLE1. Detected recruitment into the breeding
populationof web-taggedfemaleSnowGoosegoslings, tagged1973-1979.
Number
Cohort
Number
Percentage
of detected of goslings detected
recruits
a
web-tagged recruited
1973
109
3,675
3.0
1974
1975
48
124
4,471
4,443
1976
1977
1978
1979
176
192
104
63
5,440
5,306
4,861
5,448
1.1
2.8
3.2
3.6
2.1
1.2
Total
816
33,644
453
and detected recruitment rates, each hatching
period (10-12 daysin length) was divided into
three approximatelyequaltime periods(early-,
mid-, and late-hatch). Because hatching is
highly synchronized (Findlay and Cooke
1982a), the majority of goslings in any given
seasonhatch in the middle time period. The
level of synchronyis similar in all years.Table
2 is a list of the number of goslingstaggedand
the number recruited within each time period
for each gosling cohort (1973-1979). The detected recruitment
2.4
a The number of detected recruits is based on the
total number of individuals affiliated with a partic-
ular cohortwho were subsequentlyrecapturedor resightedasadultson the colonyasof 1982.The num-
rates are also illustrated
in
Fig. 1. For all cohortsexcept 1979, the numerically highestrecruitmentrateis associated
with
the earliest hatching sample,and in 4 of the 7
yr the latesthatchingsampleshowedthe lowest recruitment
rates.
Because
a multidimen-
ber is of course much lower than the absolute number
sional contingencyanalysisof recruitment by
of recruitsfrom the natal cohort in the breeding poptime period by cohort indicated a marginally
ulation.
significant three-way interaction, (P = 0.062),
suggestingthat interyear or intercohortdifferencesmay occur,eachcohortwas assessed
seprecruits.Variation among cohortsin detected arately. We detected significant differences in
recruitment rates need not necessarily repre- recruitmentratesamongthe three time periods
sent cohort-specificrecruitment, inasmuchas only for the 1973, 1974, and 1978 cohorts. In(1) birds web-taggedbefore 1976 could be de- terestingly,thesecohortsoriginated in late seatected only if they were color banded as gos- sons,(i.e. seasonsin which the mean hatch date
lings;(2) not all colorbandsareequallylegible, was later than average).The only other late
leading, for example,to lower countsfor the seasonin the sample is 1979, in which the ef1974 cohort;and (3) early cohortshave had a fect does not occur. We conclude that for some
greaternumber of seasonsduring which they seasonsat least, the highest recruitment rates
could have been detected(e.g. individuals af- are associatedwith young from the earliest
filiated with the 1979 cohort could have been
hatchingnests.
Recruitment rates depend both on survival
observedas breeding birds only in 1981 or
1982).
to reproductive age and return to the natal
To assesswhether or not there is any rela- breeding colony. Although natal philopatry is
tionship between the timing of reproduction pronounced among female Snow Geese, it is
TABLE
2. Detectedrecruitmentof goslingsfrom early-,mid-,andlate-hatchingSnowGoosenests,La Perouse
Bay, 1973-1979.
Early
Mean hatch
Mid
Re-
Late
Re-
date
1973
1974
1975
1976
1977
25 June
27 June
21 June
18 June
15 June
24
21
22
26
40
501
897
619
540
950
4.8
2.3
3.6
4.8
4.2
71
23
72
104
135
2,668
2,738
2,450
3,409
3,645
2.7
0.8
2.9
3.1
3.7
14
4
30
46
17
506
836
1,374
1,491
711
2.8
0.5
2.2
3.1
2.4
1978
1979
3 July
27 June
29
7
894
943
3.2
0.7
60
45
3,049
3,764
2.0
1.2
15
11
918
741
1.6
1.5
• *P < 0.05, ***P
cruits Tagged %
Re-
Cohort
cruits Tagged
%
cruits Tagged
%
< 0.001.
x2teststhe null hypothesisthat detectedrecruitmentisindependentof time and hatching.
X22
•
6.72*
20.8***
3.40
4.88
4.09
6.76*
2.17
454
COOKE,
FINDLAY,
ANDROCKWELL
[Auk, Vol. 101
24 1973
5.0'
4.0-
22
1975
• 1974
3.0'
21
2.0.
1.0'
E
5.0'
26
M
L
E
1976
M
L
E
M
L
1977
4O
4.0'
--
1978
135
1979
29
bJ
3.0'
17
•
2.0'
1.0.
E
M
L
E
M
RELATIVE
L
E
HATCH
M
L
E
4,5
II
M
L
DATE
Fig. 1. Detectedrecruitmentratesof early- (E), mid- (M), and late- (L) hatchinggoslings(1973-1979).
possiblethat the observedvariation in recruitment as a function of the timing of reproduction simply reflects differencesin philopatry.
Conceivably, early-hatching goslingsare more
philopatric than their late-hatching counterpartsbecauseof the greater length of time spent
in the brood-rearing area, an important factor
in the development of philopatry (see Cooke
and Abraham 1980). Alternatively, differential
recruitment may reflect true differencesin prereproductive survivorship.
The hypothesisthat differential recruitment
reflects differential survival (rather than differential philopatry) was testedusing band-recov-
recruitment of goslings from early-hatching
nestsis due to the enhanced probability of survival of these goslingsto reproductiveage.
Becauseof this variability in the age of first
breeding, one question of particular importance to the dynamics of the population is
whether or not there is any relationship between the timing of reproduction and age of
first breeding. That is, do goslingshatchedfrom
early neststend to enter the breeding population at a different age from their late-hatching
counterparts?To assess
this question,we define
N(x, t) asthe number of female goslingshatched
ery data. A large number of marked individuals are shot each year and reported to the U.S.
Fish
and
Wildlife
Service.
Most
Snow
Geese
die before reaching reproductive age, but a few
geesebanded as goslingsare shot after the age
of 3 yr (considered the average age of first
breeding). As indicated in Table 3, recovery
rates are highest among goslingsfrom the early-hatchingnests(0.48%)and lowest amongthe
goslingsfrom late-hatchingnests(0.22%).These
differencesare not significantly different at the
5% level (x22= 5.11, P < 0.1) but have the same
trend as the recruited sample.Assuming,as do
most demographers(e.g. Tanner 1978), that recoveries represent a sample of the living population, these results suggest that the higher
TABLE3. Recoveriesof web-taggedfemale goslings
shotat least3 yr after original banding.
Early
Mid
Late
Year
1973
1974
1975
1976
1977
1978
Total
Number of goslings
tagged, 1973-1978
Percentagerecovered
6
6
2
1
6
0
20
7
18
12
13
1
3
2
4
4
0
0
21
71
13
4,401
17,959
5,836
0.48
0.40
0.22
July1984]
Recruitment
andTiming
in SnowGeese
TABLE4. Recruitsfrom early-, mid-, and late-hatching goslingsin relation to age at first captureas
breeding birds.
Age
Relative
hatchdate
at first
capture
Early
2
3
17
9
4
Total
Mid
Late
Total
55
56
16
16
88
81
14
32
13
59
40
143
45
228
455
arisefrom greaterdispersalrather than higher
mortality. Higher mortality was inferred, becausea smallerproportion of the goslingswebtagged from the earlier-hatching nests appeared in the banding drives. The recruitment
data in this paper present a somewhatdifferent
picture, however. Here, the goslingsfrom the
earliest-hatching nests tended to show the
highestrecruitmentrates,implying higher survivorship. These two apparently contrasting
X22= 4.83N.S.
conclusionsmay, in part, reflect a violation of
the assumptionthat the probability of a partic-
ular family being recapturedin our banding
sample was (assumingit survived) indepenin time period t who were first recaptured as
breeding adults age x, 2 <- x <- 4. To ensure sufficient sample sizes, recruits are pooled over
cohorts
1976-1978.
If
it is assumed
that
the
probability of capturing a female on the colony
(provided that she is in fact present) is independent of the time at which she hatched as a
gosling, differencesin the age of maturation
should be manifestedin differing distributions
of age of first capture as a function of hatch
dent of its hatch date.
Conceivablythen, someof the early-hatching families considered to have suffered complete brood loss did, in fact, survive but dis-
persedout of the areain which bandingdrives
were conducted.This hypothesiscan be evaluated by consideringtwo different recruitment
groups. As mentioned previously, for seasons
1976-1979 we have two identifiable types of
recruits:(1) those caught and color banded in
date. As indicated in Table 4, there is no evitheir goslingyear and subsequentlyrecaptured
dence of such a phenomemon. We conclude or resighted as breeding adults, and (2) those
that there is no evidence that the timing of web-taggedas goslings,not capturedin their
parental reproduction affectsage of offspring goslingyear,but later caughtasbreedingadults
maturation.
and recognizedby the web-tag, which they
Findlay and Cooke (1982b) showed that ear- were still carrying. If early-hatching families
ly-hatching nests had a lower effective brood do have a lower probability of being recapsize at fledging than mid-hatching nests.This tured in our banding sample,sample 1 should
was attributed to increased gull predation on contain a smaller proportion of early-hatched
early-hatching families, who are particularly recruits than sample 2. Data in Table 5 indicate
vulnerable in the brood-rearing area. As more that for all four years (1976-1979), no differfamilies hatch,predatorswampingservesto re- enceswere detectedbetween the two samples.
duce the incidence of brood loss. Among the This suggests
that goslingsfrom early-hatching
early-hatching nests, complete brood loss is nestsdo not leavethe areamorefrequentlythan
high, whereas those families in which at least thosefrom later nestsand thusstrengthensthe
one gosling survivesto fledging have a higher conclusionsof Findlay and Cooke (1982b) that
than average brood size. Thus, disappearance mortality due to predators,rather than dispersof complete broods was cited as the explana- al, accountsfor the low fledging success
of the
tion for the lower effective brood size at fledg- early-hatching nestsin their study.
ing of early-hatching nests.
Findlay and Cooke (1982b) point out that
DISCUSSION
there is another potential explanation for the
relatively high complete brood loss from the
Our results suggest that goslings hatched
earlier-hatching nests. Only a relatively small from early neststend to have the highest resample of families is caught in the banding cruitment rates into the breeding segment of
drives, and, if earlier-hatching families have a the La Perouse Bay colony. This enhanced retendency to wander farther from the brood- cruitment probably reflects lower mortality
rearing area,they will have a lower probability amongearly-fledginggoslingsduring the preof being caught. The higher apparent disap- reproductive stagesof the life cycle. Greater
pearanceof early-hatching families might then recruitmentamongearly-hatchedyoung is most
456
COOKE,
FINDLAY,
ANDROCKWELL
[Auk,Vol.I01
TABLE
5. Comparisonof two recruitmentsamplesof
Snow Geeseto checkfor samplingbias.
early goslingsthat do survive the rigors of the
early post-hatch period (when predation appears to be heaviest) enjoy greater subsequent
Samsurvival than their later-hatchingcounterparts.
Year
plea Early Mid
Late Total
X22
Early-hatching goslings are heavier when
1976
I
15
49
18
82
2.53
caught
at banding and have longer flight feath2
I1
55
28
94
ers
(as
measured
by growth of 9th primary) than
1977
1
22
87
11
120
1.20
mid- and late-hatchinggoslings.Early-hatch2
18
48
6
72
ing goslingscaughton 30 July in 1978weighed
1978
I
3
15
3
21
2.53
1,385.9+ 19.2g, mid-hatchinggoslings1,259+
2
26
46
12
84
15.7g, and late-hatchinggoslings1,082.5+ 39.2
1979
I
4
23
3
30
2.28
g. Early-hatchinggoslingshad a 9th primary of
2
3
22
8
33
111.9 + 2.3 mm, mid-hatching goslings meaTotal
1
44
174
35
253
4.34
sured
84.2 + 1.8 mm, and late-hatching gos2
58
171
54
283
lings measured 58.4 + 5.2 mm. Because
aSample1: webtaggedas goslings,recapturedand banding is carried out as close to fledging as
bandedin goslingyear, caughtor sightedasbreeding adults.Sample2: webtaggedasgoslings,not re- possible, it seems reasonable to assume that
capturedin goslingyear, caughtor sightedasbreed- thesemorphologicaldifferencesreflect the olding adult.
er age of the earliest-hatchinggoslingsand that
in general the earliestgoslingsto hatch are the
earliestgoslingsto fledge.Thus,early-hatching
goslings undoubtedly have more time to
marked in late seasons,that is, in seasonswhere
achieve optimal condition before the fall milate snowdisappearance
leadsto a delayin re- gration. Cooch et al. (1960) showed interseaproductive activities. Abraham (1980) showed sonally that mean weights of immature geese
that mean laying date was closelycorrelated in the fall stagingareasat JamesBay in early
with date of snow disappearance
at the La Per- October were highest in seasonswith earliest
ouseBay colony. These findings lend credence mean hatch dates. They suggestedthat lower
to the views of Barry (1962) and Cooch (1961) weights at this time would increase the vulthat nesting successin arctic-breedingwater- nerability of geeseto mortality during the fall
fowl is stronglyinfluencedby the shortbreed- migration.If the sameargumentcanbe applied
ing season and that natural selection should intraseasonally,the higher recruitment of earfavor mechanismsthat result in (1) nestingas ly-hatching goslingscould be understood.
early as possiblewithin a breedingseasonand
An alternative (but not mutually exclusive)
(2) finely tuned responsesto the immediateen- explanation for the higher recruitment of earvironmental conditions on the breeding ly-hatching goslingsrelatesto parental condigrounds.In general, the earliest-layingbirds tion. Ankney and Macinnes (1978) showed a
hatch first and the latest-layinglast, but hatch- correlation between clutch size and the nuing is more synchronizedthan laying (Findlay trient-reservestatusof arriving females.There
and Cooke 1982a). Findlay and Cooke (1982b) is also a negative correlation between clutch
showed that the earliest-layingand hatching size and laying date (Findlay and Cooke 1982a).
birds had the largestclutch sizesand largest Thus,earliest-layingfemalesappearto be those
number of goslingsleaving the nest. Despite with the largestnutrient reserves.Although the
this, the earliest-hatchingnestsdid not have genetic and environmental determinants of
the highest overall fledging success:females nutrient-reserve status are unknown, a large
whoseclutcheshatchedduring the peak hatch nutrient reserve may confer a considerablereperiod had the greatest effective brood size productiveadvantage,not only in allowing for
(Findlay and Cooke 1982b)
the production of more eggsbut also allowing
In light of the findingsof Findlayand Cooke for more effective post-fledgingcare. If larger
(1982b),the enhancedrecruitmentfrom early-
nutrientreservesare indicativeof general
hatching nestsdocumentedhere is somewhat
qualityand/or condition,then perhapsit is only
high-quality birds that can nest earliest, more
effectivelycarefor their young,and ultimately
producemore offspring that survive to breed-
paradoxical.Despitefewer goslingsper nestat
fledging, more survive to be recruited into the
breeding population. This suggeststhat those
July1984]
Recruitment
andTiming
in SnowGeese
ing age. If "quality" is, like most other quantitative traits, approximately normally distributed in the population, then one would expect
only a small proportion of high-quality birds
in any one breeding season.As such, it is not
surprisingthat a relatively small proportion of
females hatch their clutchesduring the early
period. Previously,Findlay and Cooke (1982a,
1983)demonstratedhigh repeatabilityand heritability for both hatch date and clutch size,
and Hamann (1983) hasrecently shown a high
repeatability for laying date. These findings
suggestthat at least part of the variation in timing of nesting can be ascribedto additive genetic variation among females.
Findlay and Cooke(1982b)demonstratedthat
the highestreproductivesuccess
in termsof average number of goslings fledged was associated with
those females
whose clutches
hatched
457
ACKNOWLEDGMENTS
We would like to acknowledge the Natural Sciencesand EngineeringResearchCouncil, the Canadian Wildlife Service, and the Mississippi and Cen-
tral Flyway Councilsfor their continuing financial
support of the La Perouse Bay Snow Goose Project.
We thank all those who over the years have helped
to collect the voluminous data on which this paper
is based.
LITERATURE CITED
ABRAHAMß
K. F. 1980. Moult migration of Lesser
Snow Geese. Wildfowl
31: 89-93.
ANKNEY, C. D., & C. D. MACINNES. 1978. Nutrient
reserves and reproductive performance of female Lesser Snow
BARRY, t. W.
1962.
Geese. Auk
95: 459-471.
The effect of late seasons on
Atlantic Brantreproduction.J. Wildl. Mgmt. 26:
19-26.
during the middle period. This led Cooke and CARGILL, S. M., & F. COOKE. 1981. Correlation of
Findlay (1982) to suggestthat hatch date was a
layingand hatchingsequences
in clutchesof the
Lesser Snow Goose (Anser caerulescenscaerulestrait under stabilizing selection,the optimum
certs).Can. J. Zool. 59: 1201-1204.
being an intermediate date. The results describedin this paper, where recruitment is used COOCH,F. G. 1961. Ecologicalaspectsof the BlueSnow Goosecomplex. Auk 78: 72-89.
asthe measureof reproductivesuccess,
suggest
ß G. M. STIRRETT,& G. F. BOYER. 1960. Authat there may be directionalselectionfor early
tumn weights of Blue Geese(Chencaerulescens).
nesting.
Auk 77: 460-465.
It is important to remember that annual COOKE F., & K. F. ABRAHAM. 1980. Habitat and lofledging successand annual recruitment into
cality selectionin LesserSnowGeese:the role of
the breeding populationare both components
of reproductivefitnessrelatedto different stages
of the life history. We are not able at present
to measureall the components.We do not know
if timing of breeding is correlated with adult
survival rates.The high repeatability of laying
date and clutch size reported earlier does suggest, however, that annual measuresof fecun-
previous experience. Proc. 17th Intern. Ornithol.
Congr.:998-1004.
ß& C. $. FINDLAY. 1982. Polygenic variation
and stabilizingselectionin a wild populationof
Lesser Snow Geese (Anser caerulescenscaerulescens).Amer. Natur. 120: 543-547.
, C. D. MACINNES,& J.P. PREVETT.1975. Gene
flow between breeding populations of Lesser
Snow Geese. Auk 92: 493-510.
dity (fledgingsuccess
and recruitment)are rea- DARWIN, C. 1871. The descent of man and selection
sonablemeasuresof life-time fecundity,unless,
in relation to sex.London, John Murray.
of course,these are inversely correlatedwith FINDLAY,C. $., & F. COOKE. 1982a. Breeding synlongevity.
In conclusion,it is salutary to recognize that
the deductionswe make concerningthe action
of natural selectionare highly dependentupon
the
measure
of fitness
that
we
choose.
In the
chrony in the LesserSnow Goose(Ansercaerulescenscaerulescens)
I. Genetic and environmental
componentsof hatch date variability and their
effectson hatch synchrony.Evolution 36: 342351.
ß & ---.
1982b. Synchrony in the Lesser
example given in this paper, our data suggest
Snow Goose (Anser caerulescenscaerulescens).II.
that directional selectionfor early hatching is
The adaptivevalue of reproductivesynchrony.
occurringat the La PerouseBaycolony.We feel
Evolution
36: 786-799.
that recruitment into the breeding colony is a
,&
. 1983. Genetic and environmental
better measure of annual reproductive fitness
componentsof clutch size variance in a wild
than fledging success
and suggestthat selective
population of the LesserSnow Goose(Ansercaerulescertscaerulescens).
Evolution 37: 724-734.
pressuresinfluencing time of nesting are more
FINNEY,G. H., & F. COOKE. 1978. Reproductive habcomplexthan we previously envisaged.
458
COOKE,
FINDLAY,
ANDROCKWELL
its in the Snow Goose: the influence
of female
age. Condor 80: 147-158.
HAMMAN, J. 1983. Intra-seasonal clutch size reduction in Lesser Snow Geese. Unpublished M.Sc.
thesis. Kingston, Ontario, Queen's Univ.
HARRIS,M.P. 1969. Effect of laying date on chick
production in oystercatchersand herring gulls.
Brit. Birds 62: 70-75.
1966. Population studies of birds. Oxford,
Univ.
Press.
zorbills Alca torda on Skokholm.
Ibis 121: 165-
176.
KLOM?, O. 1970. The determination
of clutch size
in birds: a review. Ardea 1958: 1-124.
VAN NOORDWIIK, A. J., J. H. VAN BALEN, & W.
1981.
Genetic
Amer.
Natur.
106: 368-379.
--,
N. S. WEDD,& J. W. F. DAVIS. 1974. Mating
preferences and sexual selection in the Arctic
Skua. Heredity 33: 1-16.
PERRINS,C. g. 1965. Population fluctuations and
clutchsizein the Great Tit ParusmajorL. J.Anim.
1970. The timing of birds' breeding seasons.
Ibis 112: 242-255.
SNOW,D. W. 1958. The breeding of the blackbird
Turdus merula at Oxford.
LLOYD,C.S. 1979. Factorsaffectingbreeding of ra-
SCHARLOO.
tive rates and breeding times and its effect on
sexual selection.
Ecol. 34: 601-647.
LACK, D. 1946. Clutch and brood size in the robin.
Brit. Birds 39: 98-109, 130-135.
Oxford
[Auk,Vol. 101
variation
in the tim-
ing of reproduction in the Great Tit. Oecologia
49: 158-166.
O'DONALD,P. 1972. Natural selection of reproduc-
Ibis 100: 1-30.
SYROECHKOVSKY,
E.V. 1975. Ves yaatsi yevo vliyanie na smertnostptentsovbelych gusey(Chencaerulescens)
ha ostrovevrangelya. (Egg weight and
its effectupon mortality of nestlingsin Chencaerulescens
on the Wrangel Island.) Zoologichesky
Zhournal
54: 408-412.
TANNER,J. T. 1978. Guide to the study of animal
populations. Knoxville, Tennessee, Univ. Tennessee
Press.