The Auk
A QuarterlyJournal
of Ornithology
Vol.
109 No. 3 July 1992
The Auk 109(3):407-421, 1992
CONTINUOUS
PARTIAL
LAYING
CLUTCH
BY AMERICAN
AND
REMOVAL
COOTS
TOTAL
IN RESPONSE TO
LOSS
CLUTCH
TODD W. ARNOLD 1
Departmentof Zoology,Universityof WesternOntario,London,OntarioN6A 5B7,Canada
ABSTRACT.--Theability of American Coots (Fulica americana)to produce additional eggs
wasstudiedby experimentallyremovingsixeggsfrom the clutchduring laying (while always
maintaining at least three eggsin the nest), and by removing entire clutchesduring laying
to force birds to tenest. Coots respondedto partial (six-egg)clutch removalsby laying an
average of 11.5 total eggs, which representsa slight though significant increaseover the
mean control clutch size of 10.5. Despite this modestincreasein mean clutch size, there was
a pronouncedincreasein the proportion of cootsproducing supernormalclutches(i.e. >_13
eggs;26%of removalclutchesvs. 8% of control clutches).Egg productionwas not affected
by foodavailability,ascootswith access
to supplementalfoodwere no morelikely to respond
to partial clutch removals than were cootsfrom unsupplemented territories. In responseto
total clutch loss during laying, most coots(99/119; 83.2%) initiated a continuation clutch.
Virtually all continuationclutches(92.4%)were initiated within five days(g = 1.5 + SD of
2.2 days).Continuation clutcheswere no smaller than normal clutches,even though continuation nestershad produced several previous eggs.In 1990 and 1991, the total number of
eggsproducedby continuationnesters(all consecutivenestscombined)averaged15.8 + 8.2
and 12.8 + 5.2, respectively,comparedto 9.8 + 1.5 and 11.0 + 2.1 eggsper clutchin initial
undisturbednests.One continuationnesterproduceda remarkable35 eggsin 37 days(four
consecutivenestsplus two parasiticeggs).Neither egg size nor nestingsuccess
were reduced
amongcontinuationnesters.Collectively,thesedataprovidestrongevidenceagainstthe eggformationhypothesis,which hasbeen invokedto explainboth clutch-and egg-sizevariation
in AmericanCoots.Thesedata demonstratethat continuationnestingmay provide a better
meansthan partial clutch removalsof testingegg-formationcapabilitiesin birds.Received
5
March 1991, accepted13 January1992.
unequivocally limited by the number of folli-
of eggslaid per clutchis not determinedby the
number of initially developing follicles, but is
instead regulated by an unknown feedback
clesthatinitiatedevelopmentbeforelayingbe-
mechanism
gins. Among indeterminate layers, the number
ACCORDING
to Klomp (1970:2-3), the number
of eggslaid per clutchby determinatelayersis
that
causes additional
follicles
to
• Current address:Departmentof Biology,Univer-
developand/or previouslydevelopingfollicles
to be resorbed during the egg-laying period
(Klomp 1970:4-5).One frequentlyusedmethod
of distinguishingbetween determinateand in-
sity of Saskatchewan,Saskatoon,SaskatchewanS7N
determinate layersis to manipulate the number
0W0, Canada.
of eggspresentin the nest during egg laying
407
408
TODDW. ARNOLD
(reviewed in Kennedy 1991). Determinate layers should not alter the total number of eggs
that they lay, whereasindeterminatelayerswill
often respond to experimental egg additions or
removalsby laying fewer or more total eggs,
respectively(Cole 1917 in Klomp 1970).
The egg-formationhypothesissuggeststhat
clutch size in some speciesof birds is limited
by the ability of laying females to form additional eggs(Lack 1967,Ryder 1970,Ankney and
Macinnes 1978). Results from egg-removal experiments are often cited as evidence for or
againstthis hypothesis(Lack 1947,Winkler and
Walters 1983, Rohwer 1984, 1986, Briggs 1985,
Duncan 1986,Arnold 1990a).Clearly, if egg removalscaninducefemalesto producemoreeggs
than are found in a typical clutch, at a normal
rate of egg laying, and with no dire consequenceslater in the breeding cycle (i.e. an inability to complete incubation due to physical
exhaustion), then one could conclude that clutch
[Auk,Vol. 109
(Anasstrepera),but clutch sizein this population
averaged10.1and did not exceed12 in a sample
of 43 nests (Gates 1962:table 6). Observations
suchas this would seemto provide a powerful
rejection of the egg-formation hypothesis(see
Arnold and Rohwer1991),but continuationlaying has been recorded so infrequently among
waterfowl (and there have been so few opportunities to observecontinuationlaying) that it
would be premature to draw firm conclusions
from the few available
records.
In this paper, I use two sourcesof data to
assessthe ability of American Coots (Fulica
americana,hereafter "coots") to produce additional eggs: (1) total eggs laid in responseto
partial clutch removals;and (2) ability to produce continuation nestsfollowing total clutch
lossduring laying. Previous authors have implied that clutch sizesand/or egg sizesof coots
were limited by the availability of nutrients for
egg-laying females (Alisauskasand Ankney
size is not limited by the ability of females to 1985, Alisauskas1986, Hill 1988, 1989, Briggs
produce eggs. Few egg-removal experiments 1989). If this were true, I predicted that: (1)
have generated such clear-cut results, however. females would not respond to partial or total
Typically, femalesdo not alter their clutch size clutch removals by producing more total sein response to egg removals, or they respond quential eggsthan are found in a typical clutch;
by replacing only a few of the eggs that were or (2) if femalesdid respondby laying more
removed(reviewedin Kennedy 1991).Suchdata total eggs, they would either have to comproare ambiguous; they do not indicate whether mise the size of their eggs (e.g. Rohwer and
femaleswere incapableof laying additionaleggs Eisenhauer1989),or they would be unlikely to
(as implied by the egg-formation hypothesis), successfullycomplete incubation (e.g. Ankney
or simply unstimulatedto lay moreeggs(Klomp and Macinnes 1978).
1970). Thus, responseof a bird to egg removal
demonstratesindeterminate laying, but lack of
METHODS
responsedoesnot demonstratedeterminatelayI conducted egg-removal experiments during four
ing; it only demonstratesthat the number of
eggspresentin the nest during laying doesnot field seasons(1985-1988) on a study area near Minproximately affect clutch size (Klomp 1970, nedosa,Manitoba (50ø10'N,99ø47'W).The study area
consisted of numerous small wetlands (~0.1 to 3.0
Briggs 1985).
ha) that were inhabited by 1 to 21 pairs of breeding
Rohwer (1984, 1986)noted that laying female coots.Egg removalsstartedwith the fourth-laid egg
ducks (prairie-nesting Anas and Aythya) do not and continued daily through the ninth-laid egg; that
respond to experimental egg removalsby pro- is, six eggs were removed from each clutch, at a rate
ducing additional eggs, but a few females that
losetheir nestsduring laying will immediately
initiate a continuation nest and proceedto lay
a normal complementof eggsin the new nest,
without interrupting the normal rate of egg laying. This phenomenon has rarely been observed among waterfowl (see Rohwer 1986),
but when
it has occurred, females have laid
many more consecutiveeggsthan are ever found
in a normal clutch. For example, Gates (1962)
documented 22 consecutivedaily eggsin three
successive
nests of a marked
female
Gadwall
of one egg per day (the typical laying rate; Arnold
1990b), unless the laying female failed to produce
nine or more eggs.Hence, until they laid 11 or more
eggs, laying birds always returned to find up to 3
eggsin their nests.I beganremovalswith the fourthlaid egg becausemanipulationof nestcontentswhen
fewer eggs are present can result in high rates of
abandonment (Arnold, pers. observ.;seealsoLivezey
1980, Rohwer 1984, 1986). Coots typically have four
to six simultaneouslydevelopingfollicles (Alisauskas
and Ankney 1985, Arnold 1990b), and clutch sizesat
Minnedosaare mostcommonly8 to 12 eggs(seeTable
1). Thus, coots should have had plenty of time to
July1992]
Continuous
Laying
in Coots
409
+
u
¸
II
mr',
410
tODDW. ARNOLD
[Auk,Vol. 109
respondto egg removalsby developing additional at the time of destruction(e.g. Doty et al. 1984),so
follicles,if sucha responsewas part of their normal clutch removalswere not standardizedwith respect
nesting behavior (see also Rohwer 1986). Control to number of eggslaid. Data used in this paper are
clutcheswere nestsfound during the sametime pe- from clutchesthat I removedduring the laying stage
riod as removal clutches (clutch size declines season- (2-11 eggslaid). At completion,most initial clutches
ally in coots [Alisauskasand Ankney 1985, Arnold on my study area contained8 to 12 eggs(range 418). Thus, some cootshad undoubtedly laid the final
egg of their clutch on the day that I collectedtheir
eggs(i.e. thesebirdswere incubating,not laying),but
I had no way of assessing
whether this was the case.
Manipulations were completedbefore 1000 to allow
in 1988. In 1985 and 1986, I also conducted a small
cootsas much time as possibleto constructa new nest
number of experimentalegg additions.This involved before their next egg was due to be laid. I removed
adding six eggsto the clutch,at a rate of one to three only one clutch per cootpair, but subsequentclutches
eggsper day, so that all six eggswere addedby the sometimeswere destroyedor abandonedduring laytime the host female had laid her sixth or seventh
ing. For clutchesthat were destroyednaturally, there
was often some ambiguity as to exactly when nest
egg.
In 1987and 1988,supplementalfood was available destructiontook place.For these clutches,I assumed
to largesamplesof cootsduring egglaying,including that destructionhad occurredat the earliestpossible
birds from removaland control nests.In 1987,sup- date. Hence, estimatesof total eggslaid are minimal
plementalfood includedcrackedcornand trout chow, and estimatesof renestingintervalsare maximal.Some
which were provided independentlyin a 2 x 2 fac- cootsproducedtwo to four continuation nestsin retorial design (i.e. +corn, +trout chow). Corn is ex- sponseto repeated nest failures. These instancesof
tremely high in nitrogen-free extract (i.e. carbohy- multiple continuation nests occurred becausesome
drate, 80.4% of dry weight) and is an excellent diet cootswere extremely sensitiveto investigatordisturfor lipogenesis(Alisauskaset al. 1988),whereastrout bance and abandoned their new clutch as soon as I
chow is high (>-40%)in crudeanimal protein (Arnold discovered it. This continued to occur even after I
1990b:appendix1). Thesefood supplementswere se- ceasedflaggingnestsand marking eggs,so I believe
lected to independentlyassesslipid (corn) and pro- it was my physical presenceat the nest that caused
1990b]; hence, removal and control clutches should
have similar initiation dates).For 1985through 1987,
control nestswere visited daily to mimic the level of
investigatordisturbanceexperiencedat removal nests,
but daily visits to control nestswere not continued
tein (trout chow)limitation during egg laying (Ali-
these birds
sauskasand Ankney 1985, Drobney and Fredrickson
1985,Ankney and Afton 1988,Briggs1989,Hill 1989).
Coots readily consumedsupplementalcorn, but reluctantly consumedtrout chow. I could not determine
turbedduring experimentalclutchremovals,and often remainedintact following natural clutchdestruc-
to abandon.
Nest
bowls
were
not dis-
tion, but only 1 of 73 cootsrenestedin the samenest
bowl following total clutch loss(this occurredmore
regularly following partial clutch removal; see also
if this wasin responseto specificnutrient needs(i.e.
lipid limitation, rather than protein limitation;Drob- Fredrickson1969,Hill 1986).I assignedcontinuation
ney and Fredrickson1985,Ankney and Afton 1988), neststo particularfemales(who were not individuor due to differential palatability. But becauseadult ally banded)by a combinationof criteria,including
coots are not •ormally carnivorous (Jones 1940), I
changed the compositionof supplementalfood in
1988 in attempt to provide a more palatablediet. In
1988, supplemental food consistedof a 20:10:10:1:1
mixture of crackedcorn, layer diet for chickens(protein content >-18%), rabbit pellets (plant protein,
>-18%),oystershell (calciumsupplement),and grit
timing of nest initiation in relation to destruction of
a previous nest (i.e. a continuation nest had to have
been initiated after an earlier nest was destroyed),
proximityto a previouslydestroyednest(usually-<40
m), and mostimportantly,visualcharacteristics
of the
respectiveeggs(i.e. shell colorand spottingpattern;
Arnold 1990b:appendix2). In blind performancetri(to assist with mechanical breakdown of food in the
als to assessmy accuracyat recognizingindividual
gizzard).
coots according to characteristicsof their eggs, I
I studied continuationnesting at the samestudy achievedapproximately90%correctidentificationin
site in 1990 and 1991. In 1991, supplemental food three of four experiments(Arnold 1990b).Statistical
(samemixture as in 1988) was available to approxi- analysisof egg-sizerepeatabililty(seebelow) providmately one-half of the cootpairs.Data on continua- ed further evidencethat ][wasassigningcontinuation
tion nestingcamefrom two sources:(1) manipulative clutches to the correct individuals. Therefore, I am
data from experimental clutch removals,in which I confident that virtually all continuation nests were
removedall eggsfrom laying-stagecoot neststo in- identified correctly.In a few instances(all from 1991)
duce renesting; and (2) observational data from in which continuation nests could not have been disclutchesthat were naturally destroyedor abandoned covered becausesubsequentnest searcheswere not
during egg laying. Clutch manipulationswere con- conducted, ][ inferred that females had renested if an
ducted with the larger objectiveof determining re- unmarkednewly hatchedbroodwas observedwithin
nestingintervalsin relation to stageof the initial nest the former nestingterritory (all other broodson these
July1992]
Continuous
Laying
inCoots
wetlands had individually color-markedchicks).In
theseinstances,the brood could have been produced
by a replacementfemale,rather than the initial nesting female, but observationsfrom other territories
(for which all nestswere discoveredand egg characteristicswere compared)suggestthat replacement
females
would
411
12
ß
ß
ß
ß
ß
ß
ß
have been rare.
Three out of 73 femaleslaid eggsparasiticallyin
nestsof othercootsduring the intervalbetweende-
ß
struction
ß ß
of their
initial
nest and initiation
of their
replacementnest(seeAppendixfootnotes).In all three
cases,theseeggswere recognizedas being parasitic
on the basis of faster than normal rates of egg deposition(i.e. 2 eggs/day)in the hosts'nests(e.g.Rohwer and Freeman 1989), and they were attributed to
a particularparasiticfemale on the basisof egg characteristics(Arnold 1990b;seealsoGibbons1986,Lyon
ß
ß
ß
ß ß
ß
0
2
4
6
8
10
12
Previous clutch size
Fig. I. Renestingdelay of AmericanCootsin relation to numberof eggsin the previousclutch(Y =
0.68 + 0.16X, r 2 = 0.04, P = 0.06, n = 92).
1991, Moller and Petrie 1991). In one case the pre-
sumedparasitewasthe only other nestingfemaleon
quential eggs as any subsetof total eggs that was
producedin a singlelaying sequenceand culminated
in a completedclutch.Renestingdelay refersto the
numberof daysin which normal egg laying was delayedbeforeinitiating a replacementclutch;for insingle nest bowl in a single laying sequenceby a stance,if a coot initiated a renest on the day after
single coot (parasiticeggsthat could be recognized clutchdestruction,without missinga day of egglayby supernormallayingratesor eggcharacteristics
were ing, the renestintervalwaszero days.Cootslay eggs
excluded).A laying sequencewas definedasa series between 2400 and 0500 (Sooter 1941, Gullion 1954),
of eggsproducedby a single female in one or more well before I initiated field work. Hence, there was
nest bowls and separatedby no more than 48 h be- no dangerof overestimating
renestintervalsby one
tween consecutivelylaid eggs(laying cootstypically day due to birdshaving laid that day'seggafter I had
lay one egg per day, but laying skips occuron ap- visited their nests.
proximately3% of all potentiallaying days;Arnold
Eggsizewasdeterminedfromlinearmeasurements
1990b:table3.8). Clutch size was consideredcomplete (+0.05 mm) of length (L) and maximumbreadth(B)
if no new egg was added for two consecutivedays, using Hoyt's (1979) formula for estimatedvolume
providedthateggsin the nestwerewarmto thetouch (cm•):
(i.e. warmer than ambient temperature)and, hence,
V = 0.000507LB2.
(1)
being incubatedby at leastone parentand not simply
abandonedduring laying (abandonedclutcheswere Thismeasureof eggvolumeis highly correlatedwith
excludedin the egg-removalexperiment).Clutchsize fresheggmass(r2= 0.96;Arnold1991)andwith total
at removal nestsrefersto total eggslaid and, there- lipid, protein,andenergycontent(r2= 0.51,0.83,and
fore, includesthe six eggsthat were experimentally 0.78,respectively;
Arnold et al. 1991).Within-female
removed.At egg-additionnests,clutchsize doesnot andwithin-clutchrepeatabilitiesfor eggvolumewere
include the six experimentallyadded eggs.For the determinedusinga nestedANOVA (Lessells
and Boag
continuation-nesting
experiment,clutchsizerefersto 1987).
the numberof eggslaid in the replacementclutch,if
Statistical
analyses
were performedusingthe GLM
it was completedand subsequentlyincubated.Total and NESTED proceduresof SAS (SAS Institute Inc.
whichwerecomputed
by hand
eggsrefersto the total numberof eggslaid in one or 1985),exceptFm•-tests,
more nestsby femaleswith neststhat were destroyed (Sokaland Rohlf 1973).Means + 1 SD are presented.
duringlaying,whetheror notsuchfemalesproduced Regression
lineswere calculatedusingleastsquares.
a continuationnest,and whether or not all eggswere
partof a singlelayingsequence.
Forexample,a female
whose first clutch was destroyedat the six-eggstage
RESULTS
and who did not subsequentlyrenest would have
producedsix total eggs,whereasa femalewhosefirst
experiments.--Coots
that had six eggs
clutchwas destroyedat the seven-eggstageand who Egg-removal
experimentally
removed
from
their nests durwent on to lay a nine-egg replacementclutch (with
or withoutinterruptingthe originallayingsequence) ing egg laying producedsignificantlymore eggs
would have produced 16 total eggs. I defined se- than did unmanipulated controls (ANOVA on
the wetland, in another case there were two other
femalespresent,and in the third casethere were I1
other nesting females,but the presumedparasiteoccupied the adjacent territory.
Clutch size refersto the number of eggslaid in a
412
ToDDW. ARNOLD
[Auk, Vol. 109
pooled data, P = 0.001); however, they only
cise interval unknown), and then laid four more
replacedabout one of the six missingeggson
average(Table 1). Removal effectswere signif-
eggs.
icant in 1985 and 1986 (Table 1), and were near-
clutch size was determined
Egg-additionexperiments.--In1985 and 1986,
for six nests to which
ly significantin 1988(feedingtreatmentspooled, sixeggshad been experimentallyaddedduring
P = 0.09), but egg removal did not affect clutch the laying period;one of thesecootslaid 9 eggs,
size among any of the 1987 feeding treatments two laid 10, and three laid 11 (2 = 10.33 + 0.82).
(Table 1), or in the combined 1987 data (P =
Clutch
size of addition
nests did not differ
from
0.85). Egg removal did not affect all cootsuniformly; clutch sizesof most removal cootswere
similar to those of controls, but a significantly
larger fraction of removal cootslaid supernorContinuation
nesting.--Mostcootsresponded
mal numbersof eggs(i.e. clutchsizes>-13, 25.7% to complete clutch loss during laying by proof removal nests vs. 8.0% of control nests, Fisher
ducing replacement clutches;99 of 119 nests
control nests (1985 and 1986 combined, 11.09
+ 1.61, n = 107; F•x-test, F2,56
= 0.26, P > 0.05;
ANOVA, F•,m = 1.32, P = 0.25).
exact test, P = 0.00006, Table 1). This hetero-
geneity of responseto egg removals was reflected by among-treatment variance components.
Variance
increased
in clutch
relative
size of removal
to controls
in seven
birds
out of
eight sample groups (Table 1; sign test, P =
0.03); however, F•ax-testswere significantonly
for 1986(unsupplemented),1987(corn-fed),and
1985-1988 (pooled data). There were no significant interaction effectsbetween egg removal
and supplemental feeding for either 1987 or
1988,or for both yearscombined,so the ability
of cootsto respond to egg removal did not appear to be a function of their ability to lay more
eggs(accordingto the egg-formationhypothesis,supplementallyfed birdsshouldhavebeen
better able to lay more eggs).With data from
all four yearscombined,there was a nearly significant removal-by-year interaction effect (Table 1, P = 0.06). Hence, the absenceof a removal
(83.2%)involving 73 differentcootswere known
to be replaced (Appendix). Of the 20 clutches
that were not replaced,9 were destroyedafter
10 June, by which time almost all nest initiations had ceased(including both initial nests
and renests; Arnold 1990b). In addition, I had
no opportunity to find replacementclutchesfor
four nests in 1991 because subsequent nest
searches were not conducted
and there was no
opportunity to observepotential broods.
Most continuationnestswere initiated aspart
of the samelaying sequenceasthe earlier clutch
(65 of 92 intervals[70.7%];Appendix, including
three caseswhere females laid some parasitic
eggs).The averagerenestinginterval was 1.5 +
2.2 days.Excludingnestsfor which destruction
dateswere impreciselyknown, the averagerenesting interval was 0.8 + 1.2 days (n = 75).
Renest intervals increasedslightly with number of eggs in the previous clutch (Fig. 1), but
were unaffectedby number of previousclutches
effect in 1987 reflected a statistically different
population responseto egg removal in com- (b = 0.33, r 2 = 0.01, P = 0.31) or total number
of previous eggs (b = 0.07, r2 = 0.03, P = 0.10).
parison to the other three years.
Completed continuation nests averaged 9.1
When I included laying date asa covariatein
the preceding analyses,quantitatively similar + 3.2 eggs in 1990 (n = 23) and 11.0 + 2.3 eggs
results were obtained in all but one case; rein 1991 (n = 21; Appendix), whereascompleted
moval effectsfor 1988 becamestatisticallysig- first clutchesaveraged9.7 + 1.5 and 11.0 + 2.1
nificant when date was used as a covariate (Ta- eggs,respectively(Arnold 1990b,unpubl. data).
ble 1).Therewere significantremoval-by-laying- These meansdid not differ (1990, F•,93
= 1.60, P
date interactions in 1985, 1986, and 1988 (fed = 0.21; 1991, F•,•36
= 0.00, P = 0.95), even though
group only; Table 1). In 1985,the removal effect continuation nestswere initiated 12 and 5 days
(i.e. mean removal clutch size minus mean con- later, on average,than were 1990and 1991 first
trol clutch size) becamelarger later in the nest- clutches (1990, F•,93= 67.18, P -< 0.0001; 1991,
ing season,but during 1986 and 1988 the re- F•,•35= 22.78, P _<0.0001). In 1990, initial clutch
moval effect diminished
later in the season.
size declined by 1.0 egg over a 12-day period,
In 1988,two cootsrespondedto partial clutch whereasinitial clutch size declined by 1.2 eggs
removalsby laying additionaleggsin a second over a 5-day period in 1991 (Fig. 2). In 1990,
laying sequence.One cootlaid 10 eggs,skipped clutch size in continuation nests was nonsigfour days,and then laid 3 more eggs.Another nificantly larger than clutch size in first nests
laid nine eggs,skippedfrom 2 to 12 days(pre- after controlling for variation in laying date
July1992]
Continuous
Laying
inCoots
(Fig. 2A; mean leastsquaresfor first nests,œ=
16-
9.48; for continuation nests, œ= 9.93); in 1991,
14'
this effectbecamesignificant(Fig. 2B;meanleast
squaresfor first nests,œ= 10.9;for continuation
nests,œ= 11.9). Further analysisrevealed that
10'
the increase
in clutch size of continuation
413
12'
nests
in 1991occurredonly on fed wetlands(i.e. there
was a significantfood-by-nest-attemptinteraction, F•,•3•= 4.96, P = 0.028; mean least squares
for unfed first nests, œ= 10.6; for unfed continuation nests, œ = 10.8; for fed first nests, œ =
11.5; for fed continuation nests, œ= 13.8). This
interaction
effect does not indicate
that unfed
cootshad difficulty laying a normal-sizedcontinuation clutch, but rather that fed coots were
14
l
• 121'-..... •
able to produce exceptionally large continua-
'• 10'J
tion
+•08
clutches.
Continuation
nesters
that
initi-
ated and completedtheir replacementclutchas
part of the samelaying sequenceas their original clutch laid 13.3 + 6.1 sequential eggs in
1990(range5-27) and 16.1+ 4.3sequentialeggs
in 1991(range9-24), an averageof 3.6 and 5.1
eggsmore than controls,respectively.
Cootsthat lost their clutchesduring egg lay-
ing producedmore total eggsthan did coots
whose first clutches were undisturbed (1990
continuation nesters,15.8 + 8.2 eggs,range 235, n = 32 [data from Appendix]; 1990 undisturbed nesters, 9.7 + 1.5 [Arnold 1990b]; 1991
continuation nesters, 12.8 _+5.2, range 3-24, n
130
ß ß
_*•ß
'•'ee..ß
ßß
140
Continuation
nests
ß
"
150
1•0
1'•0
Nest initiationdate (julian)
Fig.2. Clutchsizesof initial andcontinuationnests
in relation to nest initiation dates in (A) 1990 and (B)
1991.Regressions
of clutchsize on Julian nest initiation date:(A) 1990initial clutches,Y = 21.29- 0.08X,
r 2= 0.07, P = 0.03, n = 68; 1990 continuation clutches,
Y = 23.73 - 0.10X, r 2 = 0.08, P = 0.20, n = 23. AN-
COVA comparisons:
clutcheffect,F•,9o
= 0.54,P = 0.47;
= 41 [Appendix];1991undisturbednesters,11.0 date effect,F,,90= 6.99, P = 0.01; interactioneffect,
_+2.1 [Arnold unpubl. data]).If I excludedcoots F,,89= 0.04, P = 0.84. (B) 1991 initial clutches,Y =
that never attempted to replace their initial 42.96 - 0.23X, r • = 0.23, P = 0.0001, n = 117; 1991
clutch, cootsthat I collectedduring laying for continuation clutches, Y = 15.16 - 0.03X, r• = 0.002,
other research, and coots whose territories were
not searched for subsequent nests, the mean
number of total eggsfor continuationnesters
was slightly larger (1990, 16.7 _+ 7.6; 1991,
14.2 + 5.0).In 1990,laying ratesof continuation
nestersincreasedwith total number of eggslaid
(b = 0.010, F•,27
= 5.96, P = 0.02), but declined
with number of nesting attempts(b = -0.105,
F•,27
= 8.37,P = 0.007),suggestinga significant
temporalcostassociated
with productionof nests
rather than eggs.In 1991, laying ratesof continuation nesterswere significantlyhigher than
in 1990 (F•,6•= 7.16, P = 0.01), but laying rates
were not influencedby numbersof eggsor nests
(P = 0.43 and 0.59, respectively).
Egg-sizevariation.--Meanegg volume and
P = 0.86,n = 19.ANCOVA comparisons:
clutcheffect,
F•,• = 4.24,P = 0.04;date effect,F•,• = 30.64,P =
0.0001;interactioneffect,F•,• = 2.10, P = 0.15.
(10') eggs(early,r = 0.14,P = 0.25,n = 70 nests;
late, r = 0.10, P = 0.54, n = 38 nests).
Egg size was highly repeatablebetween initial and continuation
clutches; female effects
accountedfor 62.7%of the total variation in egg
volume, whereas clutch effects accounted for
only 13.1% of this variation. Egg size was not
reducedamongcontinuationclutches(meanegg
size of second clutchesaveraged 0.47 _+ 2.07
cm• larger than in the initial clutch [representing a 1.7%increase];paired t-test, t = 1.54,P =
0.13, n = 46 pairs). There was no trade off between total eggs laid by continuation nesters
and mean egg size (Fig. 3).
clutch size were not correlatedamong removal
nests (r = 0.14, P = 0.24, n = 71 nests). This
Nest success.--Of
39 continuation
nests that
correlationremainedinsignificantif it wasbased
only on early-sequence(1-9) or late-sequence were completed and subsequently incubated,
414
TODDW. ARNOLD
1987 (Arnold 1990b;in fact, among laying females in 1987 fat and protein reserveswere 3.3
34
ß
e•32
ß
ee ee ß
ß
and 7.3 g larger, respectively),and the failure
of food supplementsto enhance the removal
ß eelee•_
I ßß
effect in 1987 and 1988 suggeststhat local food
• 28
o
ß
> 26
supply was not an important determinantof a
coot's"decision" to lay more eggs(seebelow).
I cannot explain the lack of responseto egg
ß
'','
ß
L.U 24
ß
ß
removals
22
0
5
10
15
20
[Auk, Vol. 109
25
30
35
40
Total eggs laid
Fig. 3. Relatiortshipbetween mean eõõ size and
total egg productionfor AmericanCootslosing one
or moreclutchesduring laying (datafrom Appendix).
Trend line not significant(r = 0.16, P = 0.20, n = 66).
in 1987.
Although removal cootsoften laid more total
eggs than did controls, they replaced, on av-
erage,only one of the six eggsthat had been
removed.There appearedto be pronouncedindividual variation in responseto egg removals;
somecootsproducedsupernormalclutches(e.g.
->13 eggs),whereasotherslaid asfew assix total
eggsin all but one successfullyhatched(97.4%; eggs. Individual variation in responseto egg
data exclude one nest where I collected the feremovalsapparentlywas not a function of varimale, three nests where final fate was not de-
ation in access to food resources, because coots
termined, and six "nests" that consistedonly of
brood observations),whereas apparent nest successof first nesting attempts in 1990 and 1991
receiving supplemental food were no more
(Fisher's exact test, P = 0.08).
1990b). Reid (1987) obtained similar results
likely to respondto partial clutch removalsthan
were control coots. This failure to respond to
(excluding neststhat were abandonedor de- food supplementationwas not due to poor qualstroyedduring laying, neststhat I experimen- ity or inaccessibilityof the food resources,betally destroyed,and nestswhere I collectedone causesupplementalfood did affectmany other
or both of the adults) was 86.6% (123/142). This aspectsof nesting biology (e.g. initial clutch
difference approaches statistical significance size, fledging success,growth rate; Arnold
when he suppliedsomeGlaucous-wingedGulls
(Larusglaucescens)
with supplementalfood and
DISCUSSION
removedthe first-laidegg on the day it was laid;
Egg-removal
experiments.--During
threeof four somegulls laid a fourth egg, but their propenyears(1985, 1986,and 1988,but not 1987),coots sity to do so was unrelated to supplemental
laid larger clutcheswhen six eggswere exper- feeding.
Sooter (1941) removed all but one egg from
imentally removed from their nestsduring egg
laying. When variation due to laying date was 10 American Coot nests in Iowa. Two nests were
statistically controlled, annual variation in re- immediately abandoned, but coots at the responseto experimentaleggremovalwasnearly maining eight nestswent on to lay from 14 to
significant(P = 0.06).Thus, the lack of response 18 total eggs(œ= 15.9vs. an averageclutchsize
in 1987was not a statisticalanomaly causedby of 8.6 in comparablecontrol nests).It is not clear
small samplesof nests;in fact, I obtainedthe from Sooter's (1941) account whether all eggs
most removal data in 1987. Such annual variaexceptthe first were removeddaily aslaid, and
tion in responseto egg removalsmight be pre- whether all 14 to 18 eggswere laid as part of a
dicted if cootshad unusually low food supplies single uninterrupted laying sequence.Howin 1987, or if they arrived on the breeding ever, if both these conditions were met then it
grounds with smaller lipid reservesthan nor- appearsthat his populationwas somehowbetmal (Alisauskasand Ankney 1985) and, hence, ter able to respond to egg removals, or that
were unableto produceasmanyeggsasin other removals beginning with the secondegg are
years.Although mean clutch size was smaller more likely to be successful.
In European Kestrels (Falcotinnunculus),fein 1987 than in the preceding two years,it was
somewhatlarger than in 1988,when an effect males will lay additional eggs only when reof egg removalwas detected.Cootsdid not ar- movals are initiated before incubation constanrive with larger lipid reservesin 1988 versus cy reachesabout 50% (Beukeboomet al. 1988,
July1992]
Continuous
Layingin Coots
415
see also Meijer 1990).Meijer et al. (1990) hy- and among American Cootsin responseto expothesizedthat the rising prolactin levels as- perimental reductionof clutch size during insociated with
onset of incubation
are also re-
sponsiblefor suppressingthe developmentof
additionalfolliclesand, hence,prolactinproxiinately determines clutch size. It is not known
to whatextentfemalecootsincubateduringegg
laying (the issueis complicatedby biparental
incubation),or what happensto prolactinlevels
during the egg-layingperiod. However, incubationconstancy
increases
steadilybetweenday
3 and day 6 of egg laying, as determinedfrom
handling eggsand assessing
their relative tem-
cubation (Fredrickson 1969), as well as at unmanipulated coot nests(Gullion 1954, Bett 1983,
Hill 1986).
Results from the egg-removal experiments
provided mixed support for the egg-formation
hypothesisin AmericanCoots.Although some
cootsrespondedto the removalsby laying more
eggsthan they normally would have,mostcoots
appearedunable or unwilling to replacethe six
eggs that had been removed. These results are
similar to mostother studieson determinacyof
perature (Bett 1983, Arnold unpubl. data), so egg laying in precocialand semiprecocialbirds,
perhapsinitiating egg removalswith the fourth- in which birds either failed to respondto egg
laid egg was not early enoughto affectclutch removalsor only respondedin part (McAllister
size. Consistentwith Meijer's view, removals 1958,Barry 1962,Parsons1976,Fugle and Rothwere more effective in 1985-1986
than in 1987-
1988, when natural clutch sizes were about two
stein 1977, Rohwer 1984, Winkler 1985, Reid
1987, Arnold 1990a; but for an alternative view
eggs larger (see Table 1). Thus, in 1985-1986, see Kennedy 1991). When interpreting these
cootswould have had two more days, on av- studies, however, the distinction between beerage, to respondto egg removals.The seasonal ing unable and unwilling to continuelaying
decline in responsiveness
to egg removals,as becomescrucial(Klomp 1970).If clutchsize(toevidencedby the negativeinteractioneffectbe- tal eggsto be laid) hasbeen determinedby the
tween egg removal and laying date in 1986 and time that egg removals begin, then removals
1988, also is consistentwith this hypothesis. cannot test whether or not clutch size is limited
Clutch size declinesseasonallyand onsetof in- by the ability of femalesto form eggs(Klomp
cubationincreasesseasonally(Arnold 1990b, 1970). Problemswith timing of egg removals
unpubl. data), so cootstherefore have lesstime arenottheonly methodological
concerns.
Daily
to respond to egg removals later in the season. egg removalsare designedto fool a laying feOther birds also appear to exhibit a seasonal male into "thinking" she has laid fewer eggs
declinein responsiveness
to eggremovals(e.g. than sheactuallyhas,but somefemalesmight
House Wrens [Troglodytes
aedon],Kendeigh et recognizetheseremovalsand interpret them as
al. 1956, Kennedy and Power 1990; captive partial clutch predation. If such were the case,
AmericanKestrels[Falcosparverius],
Porter1975; andif single-eggpredatorswerelikely to return
Herring Gulls [Larusargentatus],
Parsons1976; to rob the remainder of the clutch, then the
European Kestrels,Beukeboomet al. 1988). Data adaptive responseto a perception of partial
on continuation nesting in coots would seem clutch reduction might be to abandon the curto invalidate the incubation/prolactinhypoth- rent nestand begin a new nestelsewhere(e.g.
esisof Meijer et al. (1990)becausemanybirds Hall 1987a,Armstrong and Robertson1988).In
renestedimmediatelydespitelosingclutchesof this scenario,egg-removalexperimentscannot
6 to 11 eggs (Appendix); however, evidence test the egg-formationhypothesisand will not
suggeststhat prolactin levels may decline pre- indicate whether birds are determinate or incipitously following partial or total clutch loss determinate layers. Proximate physiological
(Hall 1987a, b).
mechanismsof clutch-size determination (and
Two cootsrespondedto eggremovalsby lay- when these occur) need to be better understood
ing additionaleggs,but only after severaldays in order to properly interpret egg-removalexof delay.Theseinstancesdo not representin- periments(Meijer 1990).Fortunately,studiesof
determinatelaying, but are insteadinstancesof
double clutching in the same nest bowl. Such
continuation nesting may help to resolve this
issue in at least some speciesof birds.
behaviorhasbeenobservedpreviouslyamong
Egg-additionexperiments.--Coots
did not re-
HouseSparrows(Passer
domesticus)
in response spondto experimentaleggadditionsby laying
to egg removalsduring laying (Anderson1989), fewer eggs.Although my sampleof addition
416
TODDW. ARNOLD
[Auk,Vol. 109
nestswas small (n = 6), my chancesof detecting 8 eggs in two previous clutches(Appendix).
reductionsin clutchsizeof 1, 2, 3, 4, 5, or 6 eggs Laying rateswere not reducedamongcootslaywere 0.24, 0.70, 0.94, 0.996, 0.9999, and 1.00,
ing large numbersof eggs;in fact, theseindirespectively (Sokal and Rohlf 1973; one-tailed viduals had significantlyfasterratesof egg protests).Some authorshave argued that it would duction in 1990. In contrast to these examples
be adaptive for birds to reduce their clutch size of excessiveegg production, a few coots laid
in response to intraspecific brood parasitism very small continuation clutches, and their
(Andersson 1984, Power et al. 1989; but see Roh- combined total egg production was approxiwer and Freeman1989).Although cootsare reg- matelyone "normal" clutchof eggs(e.g.female
ularly parasitizedby conspecifics
(Lyon 1991, 4094 laid four eggsin her first nest and only
Arnold unpubl. data), and there is evidencethat five eggs in a continuation nest; Appendix).
survival is reducedamonglarge broods(Lyon Other coots had long renesting intervals (->6
1991), cootsdid not respondto experimentally days; e.g. coots 4036, 4024, 4074, 4132, 56, and
enlargedclutchesby laying fewer eggs.This is 140; Appendix) or experienced laying skips
consistent with most other experiments that while producingthe continuationclutch(coots
have addedeggsto clutchesduring laying (re- 4005 and 4024). Collectively, thesedataon conviewed by Kennedy 1991; but see Andersson tinuation nesting suggestthat most cootshad
and Eriksson 1982).
little difficulty in producingmorethan a normal
Continuation
nesting.--In 1990,over 80%of all complement of eggs, and that clutch size in
coot nests that were lost during laying were mostcootsis unlikely to be limited by the abilsubsequentlyreplacedwith continuationnests. ity to form eggs (cf. Alisauskasand Ankney
Only 7 of 92 renesting intervals (7.6%) were 1985, Briggs 1989). However, a few coots apgreater than five days, which represents the parentlyencountereddifficultieswhile producminimum amount of time necessaryto develop ing replacementclutches,and for these coots
new eggsfrom unenlargedova (Alisauskasand egg-formationcostsmay limit initial clutchsize
Ankney 1985,Arnold 1990b).Hence, mostcoots or renesting potential. Thus, I reject the egginitiated continuation nestsby using follicles formation hypothesisasa single or primary exthat were already developing.All long delays planation of clutch-sizelimitation in American
(i.e. >5 days) followed nonmanipulative clutch Coots,although it may be important for some
losses, so that exact intervals between clutches
small portion of the population.
were unknown. The majorityof renestintervals
Resultsfrom my study suggestthat continwere zero or one day. The length of renestin- uationnestingis probablya bettermethodthan
tervals increasedslightly with number of re- partial clutchremovalfor testingenergeticlimnestingattempts(one day for every three nests), its on egg formation in coots.Rohwer (1986)
numberof eggslaid in the previousclutch(one arrived at a similar conclusionafter comparing
day for every six eggs),and total number of results from his own experiments on partial
previouslylaid eggs(one day for every 14eggs). clutchremovalsin prairie dabblingducks(RohAlthough this may provide evidencethat coots wer 1984) with anecdotal reports of continuawere becoming"exhausted"from continuation tion nestingin wild ducks.Continuationnestlaying, the effectsizes,nevertheless,are quite ing mayprovidean alternativemethodof testing
small.
the egg-formation abilities of some other speSome coots produced phenomenal numbers cies of birds. For experiments on continuation
of consecutiveor near-consecutiveeggs:female nestingto be effective,birds mustlay relatively
4029 laid 35 eggsin 37 days(four different nests large clutches,they must have accessto alterplus two parasiticeggs);female 4018 laid 34 native nest sites,nest construction must be rapeggs in 39 days (five nestsplus three parasitic id, and females must have ready accessto mates
eggs);and female 4009 laid 27 eggsin 28 days displayingappropriatereproductivebehavior.
(three nests;Appendix). Other cootsproduced If birds lay small clutcheswith little time bevery largereplacementclutchesfollowinga long tween successiveeggs,and if follicles require
laying seriesin oneor moreearliernests:female severaldaysto develop,then continuationnest4088laid a 17-eggclutchafter laying 3 eggsin ing cannot occurrapidly unlessbirds develop
a previousnest;female4029laid a 14-eggclutch supernumerary follicles, perhaps in anticipaafter laying 21 eggsin three previousclutches; tion of possiblenest failure during the laying
and female222 laid a 15-eggclutchafter laying cycle. Although somebirds do develop super-
July1992]
Continuous
Laying
inCoots
417
numerary follicles (e.g. Hamann et al. 1986,Ar- total clutch removals. Most previous tests of
nold unpubl. data;but seeAstheimer 1986:fig. clutch size versusegg-sizetrade offshave relied
5), it is not clear whether this represents an on observationaldata comparing different speadaptationto possiblenestdestruction.The long cies or comparing individuals within a popuegg-formationperiodsand small clutch sizesof lation (e.g. Lack 1967, Rohwer 1988, Rohwer
most seabirdsmake them unlikely candidates and Eisenhauer1989). Such testsmay not be
for rapid continuation nesting (e.g. Astheimer
valid, becausedifferent species(or individuals)
1986, Hatchwell and Pellatt 1990). Birds that
may vary markedlyin their accessto resources.
nestin cavitiesor in densecoloniesare unlikely Thosewith abundantresourcesmay invest more
to have access to alternative
nest sites and,
heavily in egg size and clutch size, whereas
therefore,it is noteworthythat manyof the best thosewith few resourcesmay have to produce
examplesof indeterminate laying obtained via small eggsand small clutches.The result would
partial clutch removalshave involved colonial- be a positive correlation between the two traits,
or cavity-nestingbirds (e.g. Northern Flicker even though each was limited by food avail[Colaptesauratus],Phillips 1887; House Wren, ability and, thus, by definition, part of a trade
Kendeigh et al. 1956,Kennedy and Power 1990; off. A better way to test for trade offs is to maWryneck [Jynx torquilla],references in Davis nipulate one variable and then assessvariation
1955; American and European kestrels, Porter in the second variable (Gustafsson and Suth1975, Beukeboomet al. 1988;Herring Gull, Par- erland 1988). In this study, I was able to masons 1976; Adelie Penguin [Pygoscelis
adeliae], nipulate the total number of eggsthat individAstheimer and Grau 1985). Open-nesting pas- ual cootslaid and, yet, this had no discernable
serinesseemto require severaldaysto construct influence on egg size (or, vice versa, egg size
a new nest (e.g. Nolan 1978) and, not surpris- had no influence on the number of eggs that
ingly, mostindividualslosingnestsduring lay- cootswere able to lay under experimentalconing experienceseveraldaysof delay before ini- ditions). Other studies have also demonstrated
tiating a new laying cycle(Scottet al. 1987and no net decline in egg size in response to exreferences therein). The small clutch size of most perimental manipulation of clutch size via egg
open-nesting passerines(i.e. 2-4), relative to removals (Kendeigh et al. 1956, Parsons 1976,
egg-formationperiods(ca. four days;Scottand Fugle and Rothstein 1977,Astheimer and Grau
Ankney 1983,Krementzand Ankney 1986),may 1985, Astheimer 1986).
also preclude rapid continuation nesting. ConThere was no evidence that birds laying adtinuation nestingis perhapsmostlikely to occur ditional eggs in continuation nestshad lower
among solitary ground- or overwater-nesting nesting success;in fact, their nest successwas
specieswith relatively large clutches(e.g. some probably higher than that of initial nesters(97
Podicipediformes,Fergusonand Sealy 1983,Arvs. 87%,respectively;P = 0.08). Becausethe two
nold pers. observ.;Anseriformes, Rohwer 1986; groups of nests exhibited minimal temporal
Galliformes, Maxson 1977, K. Martin and S. J. overlap, the difference between the two groups
Hannon pers.comm.;Rallidae,this study).Con- could be attributable to a seasonal increase in
tinuation nesting has been observed among nest success.It is clear that the data do not supLapwings(Vanellusvanellus)that lost their orig- port the idea of an intraseasonalcostof reproinal nests at the one-egg stage (Klomp 1970); duction associatedwith excessiveegg producthesebirds have proceededimmediately to lay tion. I cannot assesspossible coststhat might
a four-egg replacement clutch, thus producing have occurredduring the brood-rearing or postfive sequentialeggsat a normal rate of laying. breeding seasons,but such delayed costsseem
Egg-layingbehavior analogousto continuation unlikely given that there were apparently no
nesting has also been observed among inter- immediate costsof extended laying.
and intraspecific brood parasites (Scott and
Ankney 1983,Gibbons 1986, Lyon 1991, J. M.
Eadie, pers. comm., M. D. Sorenson, pers.
comm.),and amongbirdswith serialpolyandry
I wish to thank T. Armstrong,C. Found,P. Martin,
(Lank et al. 1985).
Egg-sizevariationand nestsuccess.--Eggsize
K. Mawhinney, and J. Morton for nest-searching
assistance.Financialsupportwasprovidedby: the Delta
wasnot markedlysmalleramongcootsthat laid
large numbersof eggsin responseto partial or
M. Chapman,Herbert and BettyCarnes,JohnK. Coo-
Waterfowl and Wetlands Research Station; the Frank
418
TODDW. ARNOLD
[Auk,Vol. 109
formation
costs limit clutch size in waterfowl?
A
per, and JosselynVan Tyne memorial funds; Sigma
Xi; the Ontario Graduate Scholarshipprogram; the
skeptical view. Condor 93:1032-1038.
University of Western Ontario Summer Research ASTHEIMER,
L. B. 1986. Egg formation in Cassin's
Auklet. Auk 103:682-693.
Scholarshipprogram; and the Natural Sciencesand
Engineering ResearchCouncil of Canada(through T. ASTHEIMER,
L. B.,ANDC. R. GRAU.1985. The timing
W. Arnold, C. D. Ankney, and S. J. Hannon). I thank
and energetic consequencesof egg formation in
W. L. Hill and B. E. Lyon for providing the impetus
the Ad61iePenguin. Condor 87:256-268.
to study continuationnesting in coots.T. Armstrong, BARRY,T.W. 1962. Effect of late seasons on Atlantic
S. V. Briggs,and F. C. Rohwer provided helpful comBrant reproduction.J. Wildl. Manage. 26:19-26.
ments on the manuscript.
BET'r,T. A. 1983. Influences of habitat composition
on the breeding ecologyof the American Coot
(Fulica americana).M.S. thesis, Univ. Wisconsin,
LITERATURE CrrED
Oshkosh, Wisconsin.
At.ISAUSKAS,
R.T. 1986. Variation in the composition BEUKEBOOM,L., C. DIJKSTRA,S. DAAN, AND T. MEIIER.
of the eggsand chicksof American Coots.Condor
1988. Seasonalityof clutchsizedeterminationin
88:84-90.
the Kestrel Falcotinnunculus:
An experimental apAt. ISAUSKAS,R. T.,AND C. D. ANKNEY. 1985. Nutrient
proach.Ornis Scand.19:41-48.
reservesand the energeticsof reproductionin BRIGGS,
S.V. 1985. Patternsof egg-laying in prairie
American
Coots. Auk 102:133-144.
ducks. Auk 102:896-897.
ALISAUSKAS,R. T., C. D. ANKNEY, AND E. E. KLAAS.
1988.
Winter
diets and nutrition
of midconti-
nental LesserSnow Geese.J. Wildl. Manage. 52:
403-414.
ANDERSON, T. R.
1989.
Determinate
vs. indetermi-
nate laying in the House Sparrow. Auk 106:730732.
ANDERSSON,
M. 1984. Broodparasitismwithin species.Pages195-228 in Producersand scroungers:
Strategiesof exploitation and parasitism(C. J.
Barnard, Ed.). Croom Helm, London.
ANDERSSON,M., AND M. O. G. ERIKSSON. 1982. Nest
BRIGGS,S. V.
1989. Food addition, clutch size, and
the timing of laying in American Coots.Condor
91:493-494.
DAVIS,D.E. 1955. Determinatelaying in Barn Swallows and Black-billedMagpies.Condor 57:81-87.
DOTY, H. A., D. L. TRAUGER,AND J. R. SERIE. 1984.
Renestingby Canvasbacks
in southwesternManitoba. J. Wildl. Manage. 48:581-584.
DRO!)NEY, R. D., AND L. H. FREDRIOCSON. 1985.
Pro-
tein acquisition:A possibleproximatefactorlimiting clutch size in Wood Ducks. Wildfowl 36:
122-128.
parasitismin GoldeneyesBucephala
clangula:
Some DUNCAN, D.C. 1986. Does food limit clutch size in
evolutionary aspects.Am. Nat. 120:1-16.
prairie ducks?Auk 103:637-638.
ANKNEY,C. D., ANDA.D. AFTON. 1988. Bioenergetics FERGUSON,
R. S., AND S. G. SEALY. 1983. Breeding
of breeding Northern Shovelers:Diet, nutrient
ecologyof the Horned Grebe,Podiceps
auritus,in
reserves, clutch size, and incubation. Condor 90:
southwestern
459-472.
408.
ANKNEY, C. D., AND C. D. MACINNES. 1978. Nutrient
FP,
m)PaCICSON,
L.H.
reservesand reproductive performanceof female
clutch
Lesser Snow Geese. Auk
550.
95:459-471.
ARMSTRONG,
T., AND R. J. ROBERTSON.1988. Parental
investment
based on clutch value: Nest desertion
in responseto partial clutch loss in dabbling
ducks. Anim.
Behav.
36:941-943.
ARNOLD,T.W. 1990a. Determinacyof clutchsize in
Horned and Pied-billed grebes.Wilson Bull. 102:
336-338.
ARNOLD,T.W. 1990b. Foodlimitation and the adaptive significanceof clutch size in American Coots
(Fulica americana).Ph.D. dissertation, Univ. Western Ontario, London, Ontario.
ARNOLD,T. W. 1991. Intraclutch variation in egg
size of American
Coots. Condor 93:19-27.
ARNOLD, T. W., R. T. ALISAUSKAS,AND C. D. ANKNEY.
1991. Egg compositionof American Cootsin relation to habitat, year, laying date, clutch size,
and supplementalfeeding. Auk 108:532-547.
ARNOLD,T. W., AND F. C. ROHWER.1991. Do egg
Manitoba.
Can. Field-Nat.
97:401-
1969. An experimental study of
size of the American
Coot.
Auk
86:541-
FUGLE, G. N., AND S. I. ROTHSTEIN. 1977. Clutch size
determination, egg size, and eggshell thickness
in the Pied-billed
Grebe. Auk 94:371-373.
GATES,
J.M. 1962. Breedingbiology of the Gadwall
in northern
Utah. Wilson
Bull. 74:43-67.
GInnONS,J. W. 1986. Brood parasitism and cooperative nesting in the Moorhen, Gallinulachloropus.
Behav.
Ecol. Sociobiol.
19:221-232.
GULLION,G. W. 1954. The reproductive cycle of
American
Coots in California.
Auk 71:366-412.
GUSTAFSSON,L., AND W. J. SUTHERLAND. 1988. The
costsof reproductionin the CollaredFlycatcher
Ficedulahypoleuca.
Nature (Lond.) 335:813-815.
HALL,M.R. 1987a. Nesting success
in Mallards after
partial clutchlossby predators.J.Wildl. Manage.
51:530-533.
HALL, M.R. 1987b. External stimuli affecting incubation behavior and prolactin secretion in the
July1992]
Continuous
Laying
in Coots
duck (Anasplatyrhynchos).
Horm. Behav.21:269287.
HAMANN, J., B. ANDPEWS,ANt) F. COOKE. 1986. The
role of follicular
atresia
in inter-
419
MAXSON,
S.J. 1977. Activity patternsof femaleRuffed
Grouseduring the breeding season.Wilson Bull.
89:439-455.
and intra-sea-
MCALLISTER,
N.M. 1958. Courtship, hostile behavsonal clutch size variation in Lesser Snow Geese
ior, nest-establishment
and egg laying in the Eared
(Anser caerulescens
caerulescens).
J. Anim. Ecol. 55:
Grebe (Podiceps
caspicus).
Auk 75:290-311.
481-489.
MEIJER,
T. 1990. Incubationdevelopmentand clutch
size in the Starling. Ornis Scand.21:163-168.
HATCHWELL,
B.J., ANDJ. PELLATT.1990. Intraspecific
T., $. DAAN,ANDM. HALL. 1990. Family planvariationin eggcompositionand yolk formation MEIJER,
in the Common Guillemot (Uria aalge).J. Zool.
ning in the Kestrel (Falcotinnunculus):
The prox(Lond.) 220:279-286.
imate control of covariation of laying date and
clutch size. Behaviour
114:117-136.
HILL, W.L. 1986. Clutch overlap in American Coots.
Condor
88:96-97.
MOLLER, A. P., AND M. PETRIE. 1991. Evolution of
1988. The effect of food abundance on
intraspecificvariability in birds' eggs: Is intraspecificnestparasitismthe selectiveagent?Pages
1041-1048in Acta XX CongressusInternationalis
Ornithologici.Christchurch,New Zealand, 1990.
New Zealand Ornithol. Congr. Trust Board,Wellington.
NOLAN,V., JR. 1978. Ecologyand behavior of the
HILL, W. L.
the reproductive patterns of coots.Condor 90:
324-331.
HILL, W.L. 1989. Reply to Briggs:The roles of endogenousand exogenousnutrient supplies.Condor 91:494-495.
Ho•r, D. F. 1979. Practical methods of estimating
volume and fresh weight of bird eggs. Auk 96:
73-77.
JONES,J. C. 1940. Food habits of the American Coot
Prairie
Warbler
Dendroica
discolor.
Ornithol.
Monogr. 26.
PARSONS,
J. 1976. Factors determining the number
with notes on distribution. U.S. Dep. Interior
and size of eggslaid by Herring Gulls. Condor
Wildl.
78:481-497.
Res. Bull. No. 2.
KENDEIGH, S.C., T. C. KRAMER, AND F. HAMERSTROM.
1956. Variations in egg characteristicsof the
House
Wren.
KENNEDY, E. D.
Auk 73:42-65.
1991.
Determinate
and indetermi-
nate egg-laying patterns:A review. Condor 93:
106-124.
KENNEDY,E. D., AND H. W. POWER.1990. Experiments on indeterminate laying in House Wrens
and EuropeanStarlings.Condor 92:861-865.
KLOMP, H. 1970. The determination
of clutch size
in birds, a review. Ardea 58:1-125.
KREMENTZ, D. G., AND C. D. ANKNEY. 1986. Bioen-
ergeticsof eggproductionby femaleHouseSparrows. Auk
103:299-305.
I and II. Ibis 89:302-352.
LACK, D. 1967. The significance of clutch-size in
Wildfowl
19:125-128.
LANK, D. B., L. W. ORING, AND S. J. MAXSON. 1985.
Mate and nutrient limitation of egg-laying in a
polyandrousshorebird.Ecology66:1513-1524.
LESSELLS,
C. M., ANDP. T. BOAG.1987. Unrepeatable
repeatabilities:A commonmistake.Auk 104:116121.
LIVEZEY, B.C.
laptesauratus.Auk 4:346.
PORTER,R. D. 1975. Experimental alterations of
clutch-sizeof captive American KestrelsFalco
sparverius.Ibis 117:510-515.
POWER, H. W., E. D. KENNEDY, L. C. ROMAGNANO, M.
P. LOMBARDO, A. S. HOFFENBERG,P. C. STOUFFER,
AND T. R. McGuIRE. 1989. The parasitism insurancehypothesis:Why Starlings leave space
for parasiticeggs.Condor 91:753-765.
REID, W.V.
1987.
1980. Effects of selected observer-re-
lated factors on fates of duck nests. Wildl.
Soc.
Bull. 8:123-128.
LYON,B.E. 1991. Broodparasitismin AmericanCoots:
Avoiding the constraintsof parental care. Pages
1023-1030 in Acta XX CongressusInternationalis
Ornithologici. Christchurch,New Zealand, 1990.
New Zealand Ornithol. Congr. Trust Board,Wellington.
Constraints
on clutch size in the
Glaucous-wingedGull. Stud.Avian BioL 10:8-25.
ROHWER,
F.C. 1984. Patternsof egg laying in prairie
ducks. Auk
LACK,D. 1947. The significanceof clutch-size.Parts
waterfowl.
PHILLIPS,
C.L. 1887. Egg-layingextraordinaryin Co-
101:603-605.
ROHWER,
F.C. 1986. Responseto D.C. Duncan.Auk
103:638-639.
ROHWER,F. C. 1988. Inter- and intraspecific rela-
tionshipsbetween egg size and clutch size in
waterfowl.
Auk
105:161-176.
ROHWER,
F. C., ANDD. I. EISENHAUER.
1989. Egg mass
and clutchsizerelationshipsin geese,eiders,and
swans. Ornis
Scand. 20:43-48.
ROI-IWER, F. C., AND S. FREEMAN. 1989.
The distri-
bution of conspecific
nestparasitismin birds.Can.
J. Zool. 67:239-253.
RYDER,
J.P. 1970. A possiblefactorin the evolution
of clutch size in Ross' Goose. Wilson
Bull. 82:5-
13.
SASINsTITUTE.1985. SASuser'sguide: Statistics,version 5 ed. SAS Institute Inc., Cary, North Carolina.
SCOTT,D. M., AND C. D. ANKNEY. 1983. The laying
420
TODDW. ARNOLD
American
Coot, Fulica americana americana Gmelin. Ph.D. dissertation, Iowa State Univ., Ames.
cycle of Brown-headed Cowbirds: Passerine
chickens?
Auk
100:583-592.
SCOTT,D. M., R. E. LEMON, AND J. A. DARLEY. 1987.
Relaying interval after nest failure in Gray Catbirds and Northern
Cardinals.
Wilson
[Auk,Vol. 109
WINKLER,
D.W. 1985. Factorsdetermininga clutch
size reduction in California Gulls (Laruscalifor-
Bull. 99:
nicus):A multi-hypothesis approach. Evolution
SOKAL,R. R., AND F. J. ROHLF. 1973. Introduction to
biostatistics, 2nd ed. W. H. Freeman, San Fran-
WINKLER,D. W., AND J. R. WALTERS. 1983. The de-
39:667-677.
708-712.
termination of clutchsize in precocialbirds. Curr.
Ornithol.
cisco.
1:33-68.
SOOTER,
C.A. 1941. Ecologyand managementof the
APPENDIX.Layinghistoriesof AmericanCootsexperiencingtotal clutchlossesduring egg laying.a
Mean
Totals
No. eggslaid in clutch
egg
Interval (days)between clutches
3-4
4-5
5-6
Lay
Female
1
2
3
4
5
1-2
2-3
4029
5
7
7
14 b
--
2 ca
2
0
--
--
35 d
4018
6
6
7
4
8
3'
0
4
1
*
34'
4009
3
10
14
0
!
--
--
--
27
4036
10f
8
4
5
--
8fs
3
3
--
4026
3
7
1h
6
9
0
7h
1
2
--
4022
1
9
13
--
0
1
--
--
4015
7
7
7
1
0
--
4005
11
9
--
1.....
4088
3
16
--
0
....
4024
8f
11
--
6 fi
....
4074
3
0
11•
1f
....
Eggs
Days
size
rate
Nests
37
0.946
39
0.872
28
27+
26+
---
Nest
(cm')
success
4
28.26
Hatch
5
31.18
--
0.964
3
27.88
Hatch
4!
0.659
4
32.61
--
36
0.722
5
31.84
Hatch
23
24
0.958
3
30.2!
Hatch
21
22
0.955
3
27.82
--
20
23'
0.870
2
25.60
Hatch
19
19
1.000
2
25.12
Hatch
19+
27+
0.704
2
29.06
Hatch
30
0.633
3
24.24
Hatch
1990
8•
4011
5•
12
4044
4
5
4006
6
10
4025
6
4012
--
8
--
--
--
8
--
--
--
19+
17+
18
0.944
2
31.58
Hatch
--
--
17
20
0.850
3
31.59
--
2
1
--
0
....
16
16
1.000
2
30.4!
Hatch
10
--
0
....
16
16
1.000
2
29.74
Hatch
6
!0
--
0
....
16
16
1.000
2
32.70
Fail
4037
8i
8
--
4
....
16
22
0.727
2
27.65
Hatch
4060
7
8
--
0
....
15
15
1.000
2
28.51
Hatch
4033
5
8
--
2
....
13
15
0.867
2
28.38
Hatch
4133
5f
8
--
3•
....
13+
16
0.813
2
29.12
Hatch
4082
5
6
--
2k
....
12 •
13
0.923
2
28.04
Hatch
4046
5
7
--
4
....
12
16
0.750
2
29.31
Hatch
4055
2
9
--
1
....
11
12
0.917
2
22.84
Hatch
4080
6
5
--
3
....
11
14
0.786
2
28.96
--•
4027
3f
7
--
1•
....
10+
11
0.909
2
27.74
Hatch
4132
4•
1f
10+
21
0.476
3
24.95
--
4094
4
5
9
10
0.900
2
27.75
Hatch
--
5
3f
-3
8•
1
....
4
0
4007
1•
3
4080
•
2
--
0
4049
2 h•
3
--
1
....
4034
2,
......
4084
2i
--
--
*
....
149
10
14
--
1
7•
15
--
1•
87
10
4
7
144
7
1
134
4
16
--
--
*
--
--
7
11
0.636
3
25.93
--
--
--
5
5
1.000
2
30.15
--
5+
6+
0.833
2
27.68
Hatch
2
2
--
1
27.57
--
2
2
--
1
29.99
--
....
24
25
0.960
2
24.23
Hatch
....
22+
23
0.957
2
31.66
Hatch
1991
222
@
--
0
0
X@
--
--
21+
21+
1.000
3
27.18
Hatch@
13
1
0
--
--
--
21
22
0.955
3
31.76
Hatch
--
0
....
20
20
1.000
2
24.92
Hatch
July1992]
Continuous
Layingin Coots
APPENDIX.
421
Continued.
Mean
Totals
No. eggslaid in clutch
Female
I
2
225
6
14
23
2
7
3
4
egg
Interval (days)betweenclutches
5
1-2
2-3
--
3
....
10
3
2
3-4
4-5
5-6
--
--
--
--
--
--
Lay
Eggs
Days
20
19
size
(cm3)
Nest
rate
Nests
success
23
0.870
2
25.01
Hatch
24
0.792
3
29.01
Hatch
18
18
L000
2
28.11
Hatch
17
17
1.000
3
31.53
Hatch
17
17
1.000
2
25.39
Hatch
46
8
10
--
0
....
147
I
5
11
0
0
203
4
13
--
0
....
196
4
3
0
0
16
16
1.000
3
30.05
Hatch
162
6
10
--
0
....
16
16
1.000
2
31.30
Hatch
48
3
12
--
0
....
15
15
1.000
2
28.58
Hatch
152
4
11
--
1
....
15
16
0.938
2
27.62
183
123
2
9
9
5
4
@
0
2
3
@
15+
14+
18+
16+
0.833
0.875
4
3
26.54
--
Hatch@
Hatch@
Hatch
9
@
--
--
X@
--
--
---
--
---
160
1n
12
--
0
....
13+
13+
1.000
2
29.97
138
9
4•
--
0
....
13+
13+
1.000
2
30.20
168
3
10
--
0
....
13
13
1.000
2
27.02
Hatch
175
1•
12
--
3f
....
13+
16
0.813
2
29.41
Hatch
12+
12+
1.000
5
28.74
Hatch@
12
13
0.923
2
25.84
__m
1.000
2
31.21
139
3
4
40
2
10
1
4
@
170
5
6....
31
5
5
I
--
--
306
2•
9
--
--
--
--
0
0
1
....
0
X@
--
0
__m
--
--
--
11+
11+
0
0
X
--
--
11
11
1.000
3
28.09
1•
....
11+
12
0.917
2
--
--
61
6
5•
--
--
--
I
....
ii+
12+
0.917
2
27.01
193
6
4
@
--
--
0
X@
--
--
--
10+
10+
1.000
3
27.24 Hatch@
--
--
--
10
10
1.000
2
25.13
10+
12+
0.833
2
__
10
18
0.556
2
28.29
--
9
9
1.000
2
26.39
Hatch
30
6
4
--
--
--
0
*
176
1•
9
--
--
--
2•
....
56
7
3
--
--
--
8
192
I
8
--
--
--
0
155
4
5
--
--
--
0
117
2
2
--
--
1
140
i t"
8
--
--
--
6•
....
171
2
6
--
--
--
2
X
93
7
........
187
6
@
199
6•
......
200
2
3
204
4•
--
142
3
......
5
--
X@
--
0
--
*
--
--
--
....
0
--
--
--
--
9
9
1.000
2
29.79
--
X
--
--
9
10
0.900
3
25.56
--
9+
15+
0.600
2
29.57
Hatch
--
--
--
8
8
1.000
2
28.28
--
7
7
--
I
--
--
6+
6+
--
2
27.11 Hatch@
6+
6+
--
I
30.11
5
5
1.000
2
27.71
--
4+
4+
--
i
--
--
3
3
--
1
--
--
....
-....
--
--
--
' Estimates
followedby + areminimal.*, noreplacement
nestfounddespite
activesearching.
X, nonestsearch
conducted.
@,unmarked
brood
later sightedon territory,presumablyfrom replacementclutch.
bComplete
(incubated)
clutches
underlined;
all otherclutches
removed,
destroyed,
or abandoned
duringlaying.
• Renestintervalsbasedon experimental
clutchremovals
areunderlined;
all othersfromnaturallydestroyed
or abandoned
nests.
dTwo eggslaid parasiticallyin anothercoot'snest.
• Three eggslaid parasiticallyin anothercoot'snest.
•Exactdayof clutchdestruction
unknown;clutchmayhavebeencomplete
and/orlargerthanindicated;
renestintervalmayhavebeenshorter
than indicated.
gMalekilledon nest(probablyby a GreatHornedOwl, Bubo
virginianus);
femalerematedfor second
clutch.
hNestdestroyed
whenfound;estimates
of eggslaidareminimums
andestimates
of renestlng
delaysaremaximums.
• Two additionallaying skipswithin laying sequencefor clutch2.
•Clutch1 wasa renestproduced
afteran earlierclutchhadbeendestroyed
duringincubation.
kOne egg laid parasiticallyin another coot'snest.
' Nestexperimentally
destroyedto inducerenesting.
= Femalecollectedfor energetics
study.
"Nest
fate not determined.