October 1991] ShortCommunications 969

Natural Environment ResearchCouncil grant GR3/ LINDEN, M., & A. P. MOLLERß 1989ß Cost of repro- 7249. duction and covariation of life history traits in birds. Trends Ecol. Evol. 4: 367-371.

LITERATURE CITED NUR, N. 1984. The consequencesof brood size for breeding Blue Tits I. Adult survival, weight ASKENIVIO,C. 1979. Reproductiverate and the return change and the cost of reproduction. J. Anim. rate of male Pied Flycatchers.Am. Nat. 114:748- Ecol. 53: 479-496ß 753. 1988a. The consequencesof brood size for COHEN,J. 1988. Statistical power analysis for the breeding Blue Tits III. Measuring the cost of re- behavioral sciences,second ed. New Jersey,Law- production: survival, future fecundity and dif- rence Erlbaum Assoc. ferential dispersalßEvolution 42: 351-362ß DESTEvEN,D. 1980. Clutch size, breeding success ß 1988b. The costof reproduction in birds: an and parental survival in the Tree Sparrow (Iri- examination of the evidenceß Ardea 76: 155-168. doprocnebicolor). Evolution 34: 278-291. --. 1990. The costof reproduction in birds, eval- DIJKSTRA,C., A. BULT, S. BIJLSMA,S. D^AN, T. MEIJER, uating the evidence from manipulative and non- • g. ZIJLSTRA.1990. Brood size manipulations manipulative studies. In Population biology of in the Kestrel (Falco tinnunculus):effects on off- passerinebirds, an integratedapproach (J. Blon- spring and parent survival. J.Anim. Ecol.59: 269- del, A. Gosler,J. D. Lebreton,and R. McCleery, 285. Eds.).Berlin, Springer-VerlagPressß FLEISS,J. L. 1981. Statistical methods for rates and ORELL,M., & K. KOIVULA. 1988. Costof reproduction: proportions, second ed. New York, J. Wiley and parental survival and production of recruits in Sons. the Willow Tit Parusmontanus. Oecologica 77: 423- FORBES,L.S. 1990. A note on statisticalpower. Auk 432. 107: 438-439. PETTIFOR,R. A., C. M. PERRINS,& R. H. McCLEERY. GUSTAFSSON,L., & W. J. SUTHERLAND.1988. The costs 1988. Individual optimization of clutch size in of reproduction in the Collared Flycatcher Fice- Great Tits. Nature 336: 160-162. dula albicollis. Nature 335: 813-815. REID,W. V. 1987. The cost of reproduction in the HARRIS,M.P. 1970. Breedingecology of the Swal- Glaucous-wingedGull. Oecologica74: 458-467ß low-tailed Gull Creagrusfurcatus. Auk 104: 470- ROSK•Fr,E. 1985. The effect of enlarged brood size 480. on the future reproductive potential of the Rook. HEGNER,R. E., & J. C. WINGFIELDß1987. Effects of J. Anim. Ecol. 54: 255-260. brood-sizemanipulation on parental investment, SOK•L, R. R., & F. J. ROHLF. 1981. Biometry, second breeding success,and reproductive endocrinol- ed. San Francisco, W. H. Freeman and Co. ogy of House Sparrows.Auk 104:470-480. TOFT,C. A., & P. J. SHEA. 1983. Detecting commu- KORPIM;(m,E. 1988. Costsof reproductionand suc- nity-wide patterns: estimating power strength- cessof manipulated broods under varying food ens statistical inferenceß Am. Nat. 122: 618-625. conditionsin Tengmalm'sOwl. J.Anim. Ecol.57: 1027-1039. Received19 September1990, accepted26 May 1991.

Courtshipof Ducklingsby Adult Male Chiloe Wigeon(Arias sibilatrix)

GWEN BREWER1 JamesFord Bell Museum of NaturalHistory, University of Minnesota,10 Church Street SE, Minneapolis,Minnesota 55455 USA

Unlike mostbirds with seasonalpair bonds,many lead them away from other birdsßWhen femalesshow waterfowlpair muchin advanceof the nextbreeding their preferencefor one male and rejectionof others season(Rohwer and Anderson1988). For example,in by giving Inciting displays,temporary associations Holarcticdabbling ducks (Anatini), courtship and pair between males and females are formed. As bonds formationoccur up to 8 monthsbefore breeding be- strengthen, the members of a pair maintain close gins (Hepp and Hair 1983).During the early period proximity, synchronize their activities, repel rivals, of bond formation,males court femalesand try to and perform displays that reinforce the bond (Mc- Kinney in press). Pair bonds continue to be tested during spring migration and after arrival on the •Present address:Department of Biology, Frost- breedinggrounds, although most birds are pairedby burgState University, Frostburg, Maryland 21532USA. this time. 970 ShortCommunications [Auk, Vol. 108

Several hypothesesexist to accountfor the early binations of these characteristics. I identified broods timing of pair formation in the Northern Hemisphere throughout study periodsby comparingthe size and (reviewedby Rohwerand Anderson 1988). Generally, age of the ducklings,the locationof the brood, and the balancebetween benefits from pairing to both the plumagecharacteristics of parents(both male and sexesand coststo malesis of major importance.Al- female attend broods)with previous information and though participatingin courtshipmay be energeti- sketchesof both parents (Brewer 1990). Ducklings cally costly to females as well as males, females are were assignedto one of seven plumage classes(Ia-c, expectedto pair as soon as possiblebecause of the IIa-c, IIIa; Gollop and Marshall 1954)and assignedto benefits(such as enhancedfeeding efficiency)that age groups on the basis of data from 15 known-age they receive when paired. Males must balance the broods(those first sighted within 1-3 days of hatch- benefitsof pairing againstthe costsof mate defense ing). I could determine the sex of ducklings ca. 8 and vigilance, so that the exact timing of pair for- weeks of age (classIIc) from wing patterns (males mation will be determined largely by the male. have darker speculaand bright white upper wing Although the courtship displays of most tropical coverts)and from the presenceof a two-syllablewhis- and SouthernHemisphere dabbling ducks have been tle given only by males (method confirmed in cap- described(Johnsgard 1965), the timing of courtship tivity for 14 ducklings,pets. obs.). Six ducklingswere and pair formation for thesespecies is poorly docu- identified by the unique pattern of down and feathers mented.In theseregions, the timing of pair formation on their heads and bodies. may be influencedby variable or irregular breeding Male courtshipactivity.--In general,male courtship seasonsand the lack of seasonalor molt migrations of ducklingswas similar to male courtshipof adult (McKinney 1985, Oring and Saylet in press).In sev- females (Johnsgard1965; pets. obs.). Males directed eral SouthAmerican species, including the migratory four courtshipdisplays at ducklings (Rakoo,Display Chiloe Wigeon (Anassibilatrix), courtship activity has Shake, Grunt-whistle,and Preen-behind-the-wing; been noted before the end of the breeding season Johnsgard1965, Brewer 1990) and performed Pre- (Weller 1968),but the significanceof this exception- copulatory Head-pumping (Johnsgard 1965) near ally early courtship is unknown. I found that un- them. Group flightswith I-2 juvenile femalesand 2- pairedmales court and form bondswith ducklingsas 8 adult males,similar to courtshipflights described part of a pair formationprocess in Chiloe Wigeon. I for adult American Wigeon (Anasamericana; Wishart know of no previous descriptionsof this behavior in 1983),occurred late in the study period. I identified any waterfowl species. the intendedreceiver of a displayby noting the body Courtshipof ducklingswas discoveredincidentally orientationof the performer(McKinney in press),and during a study of biparental care in Chiloe Wigeon I considerhere only thosedisplays that were unam- (Decemberthrough February, 1985-1988;total 500 h biguouslydirected at ducklings(60 displayswere di- of observation)at five locationsin southernArgentina rectedat paired females;no unpaired femaleswere within 70 km of San Carlos de Bariloche (41ø08'S, present). 71ø15'W)and one location in BuenosAires province For 12 known-agebroods, the age at which duck- (38ø51'S,60ø05'W). Although similar trends in court- lings were first courtedranged from 9 days(class Ib) ship were observedat all six areas,I focusedon data to 33 days (classIIa) (mean [+SD] = 21 + 6.6). The from the three main study areasin southernArgen- numberof displayingmales at one locationranged tina: Laguna Los Juncos,Estancia San Ramon, and from oneto 23,with up to sevenmales simultaneously Arroyo Pilcaniyeu (locationsand characteristicsfur- courting one or more ducklings. No male parents ther describedin McKinney and Brewer 1989). Fam- courtedtheir own ducklings,although mostpairs ac- ilies or courting maleswith ducklingswere observed companiedthem until ducklingswere at least6 weeks during 30- to 60-min periodsat all timesof day using old. Male ducklingsor juvenilesdid not courtfemales focal animal sampling (Altmann 1974) to record be- during the study period. havioral interactions and the activities of males, fe- Almostall courtshipdirected at ducklings(2,800 of males,and ducklingsevery 30 s. I observedcourtship 2,810displays) was performedby adult malesjudged during 14 days (45 h of observation)at EstanciaSan to be unpaired becausethey did not associatewith Ramon,35 days(102 h) at Arroyo Pilcaniyeu, 19 days an adult female for at least I h or exhibit behavior (22 h) at Languna Los Juncos,and 25 days (22 h) at seen when pairs are separated.(Courtship of duck- the other three locations. lings did not startuntil after mostbroods had hatched, Although individuals were not marked, all 28 pairs and it was very unlikely that paired maleswould be of parents and their broods, 15 courting males, and away from their own matesfor such long periods.) 6 eight-week-old ducklings were identified consis- These males increasedas a proportion of the local tently by morphologicalfeatures. Intensity and extent population and in number throughout the breeding of rust coloration on the flanks and under the tail, seasonat all study areas.Mean sex ratios (male: fe- amount of white on the forehead, and the shape of male) ranged from 1.26:1 (Arroyo Pilcaniyeu lAP]) whitish cheek patchesvaried in parental adults, and and 1.41:1 (Laguna Los Juncos[LLJ]) in December somecourting maleswere identified by unique corn- (mean number of adultspresent = 41 and 47) to 2.13:1 October1991] ShortCommunications 971

[] negativerespon• malesnear them. Ducklings alsogave Chin-lifts dur- (ducklingsgave chin-lifts and went to parents) ing or after parentalaggression to courtingmales (n ß positiveresponse = 214 bouts)and during family group displaysafter N=6 (duckh•s gave chin-hits and went to coualngma•e) aggression(n = 76), similar to family Triumph Cer- emonies associatedwith aggressionin geese(Hein- roth 1911). Ducklings sometimesgave DistressCalls asthey tried to returnto their parents,and some dived repeatedlywhen smallgroups of malesdisplayed near them. 5-7 8-10 Duckling age m weeks In the 5-7 and 8-10 week agecategories, some duck- lings still escortedby their parentsresponded posi- Fig. 1. Mean responsesper hour to courting males tively to courtingmales by going towardsthem and by ducklings accompaniedby their parents.Number of broods above bars. Number of broods is small for giving Chin-lifts (Fig. 1), justas paired adult females do to their mates(Lorenz 1951-1953).Up to five duck- the 2-4 week age group becausefewer ducklings of lingsshowed a preferencefor certaincourting males, this agereceived displays; in the 8-10 week agegroup, although no maleswere able to defend associations few broodswere still accompaniedby their parents. with more than one duckling for longer than 3-7 days.These males had no distinguishingphysical fea- (AP) and 3.28:1(LLJ) (• = 47 and31 adults)in January tures. and 3.0:1 (AP) in February (œ= 69 adults). To test the observed differencesin duckling be- Males directed courtshipdisplays at ducklingsof havior by age,each brood in an agegroup was scored all age classes,but courtshiprate (displaysper hour) as positiveor negativebased on whether ducklings and frequency (percent hour of observationwith respondedoverall morepositively or morenegatively courtship)increased with duckling age (2-4 weeks: towardsmales when they were in that ageclass. Using rate = 0.67-2.6, frequency= 13-34%;5-7 weeks:rate thesedata, responseswere found to depend on age = 7.2-14, frequency= 63-72%;8-10 weeks:rate = (Gaa•= 9.982,df = 2, P < 0.01;2-4 weeksold.' n = 16 9.3-15, frequency= 79-95%). In addition, malesdi- broods; 5-7 weeks: n = 16; 8-10 weeks: n = 5). At recteda higher proportion of Preen-behind-the-wing Arroyo Pilcaniyeu,where all three age classeswere displays(used in pair maintenance)and Pre-copula- observed,negative responses per hour decreasedwith tory Head-pumpingat 8-10-week-oldducklings (29%, age (4.8, 3.7, 1.1) while positive responsesincreased n = 234displays) than at 5-7-week-oldducklings (10%, in frequency(0, 1.6, 1.8). n = 77). Formationof bonds between adult males and ducklings.- Although the overall sexratio of 8-week-old duck- As ducklings began to remain with courting males lingswas approximately 1:1, adult males directed more and give Chin-lifts besidethem, bondsbegan to form courtshipdisplays at female ducklings(n = 430 dis- between unpaired adult males and 6-8-week-old plays)than at male ducklings(n = 56) and were more ducklings (classIIb-IIc). Activities of ducklings ac- aggressivetowards male (n = 42 aggressiveactions) companiedby a courtingmale or by a male parent than towardsfemale ducklings(n = 7). Courting males did not differ significantly(Mann-Whitney test, n• = were frequentlyaggressive to eachother (352 threats, 9 male-duckling "pairs," n2 = 10 families, P > 0.05), 236 chases,and 91 fights in 110 h of observation),as but courtingmales spent significantly more time feed- in other wigeon (Wishart 1983),and a greaternumber ing (Mann-Whitney test, n = 9, n2 = 10, T = 111, P of aggressiveactions occurred near female ducklings < 0.05) and less time alert (T• = -3.6, P < 0.001) or (89 actions near females, 16 near males). engagedin aggressiveinteractions (T, = -3.008, P < Responsesof ducklingsand parents.--Bothmale and 0.002)than male parentsthat accompaniedducklings female parentswere generally intolerantof courting in age classIIc. Bondsbetween ducklingsand their males near their ducklings and repeatedly threat- parentsbegan to breakwhen ducklingsleft their par- ened, chased,and peckedthem (total 2,060aggressive ents and spent time with courting males, and when actionsduring 186 h of observation).Some courting parentswere unable to keep their ducklingsclose to malesappeared to be toleratedmore than others,al- them due to the almostconstant intrusion of courting though these "favored" males did not differ in any males.In areaswith fewer courting males,bonds be- obviousway from the other males.Some ducklings tween parents and ducklings did not break until of all agesresponded negatively to courtingmales by ducklingswere 9-11 weeks old, which suggeststhat returning to their parents,running or swimming away the presenceof courting maleswas an important fac- from the males,or directing Chin-lifts--a displaysim- tor in the timing of family break-up(Brewer 1990). ilar to the Inciting of adult females(Johnsgard 1965)-- Almost all "pairs" including known-sexducklings to their parents(Fig. 1). Someparental aggression (n were composedof adult malesand female ducklings, = 16) occurredafter ducklingsgave Chin-lifts when and they behaved similarly to adult pairs of ducks the ducklings appeared to be soliciting aid and in- (e.g. CommonGoldeneyes, Bucephala clangula; Afton citing their parents to threaten and chase courting and Sayler 1982). Both individuals defended their 972 ShortCommunications [Auk, Vol. 108 partneragainst other adults and ducklings,associated ducklingsand possiblyparents influencing the out- closely,and gave pair-reinforcingdisplays (Preen- come. behind-the-wingby male adults,Chin-lifts by duck- Other factorsare necessaryto explain why court- lings)and pair palaverssimilar to thoseof adult pairs ship and bond formation occur so early in Chiloe of Chiloe Wigeon (Johnsgard1965). Ducklingses- Wigeon and to determine if the adult sex ratio is cortedby malesalso gave Chin-lifts to incite aid (n particularlyimportant. The adult sexratio washighly = 62 bouts)and after intruding males were chased biasedtoward males in my study populations.Sex away (n = 112).One male copulatedonce with the ratioswere considerablyhigher than thosenoted for femalejuvenile that he wasaccompanying (only the northern dabbling ducks,which range from 1.22:1to male gavepre- and post-copulatorydisplays). 1.56:1 (Bellrose et al. 1961). More information is need- Althoughseveral adult males accompanied and de- ed on sex and age ratios in Chiloe Wigeon, and on fendedmale ducklings,only twice wasa male duck- mortalityfactors (such as predation and energystress ling (classIIc) escortedfor an entire 1-hwatch. Youn- that maybe greateron females)that are likely to affect ger male ducklings(class IIc) behavedlike female the sex ratio (Bellrose et al. 1961). ducklingswhen adult malesaccompanied them, but The operational sex ratio (Emlen and Oring 1977) classIIIa and juvenile malestended to avoid adult may be further skewedin favor of males in Chiloe males. As the breeding seasonprogressed, female Wigeon becauseof the presenceof long-term pair ducklingsand juvenileswere conspicuouslyaccom- bonds. Data from both wild (Weller 1968,Brewer 1990) paniedby adult males,and male ducklingsand ju- and captive birds (Heinroth 1911, Kear 1970,Brewer venilestypically spent time aloneor with othermale 1990) suggestthat long-term pair bondsare charac- ducklings. teristic of this species.If population and operational Information on the persistenceof adult male-fe- sexratios are highly male-biased,and if birds remain maleduckling "pairs" was limited to 15cases in which in long-term bonds with little mate-switching,few I could identify the male--and in some casesthe femalesexcept young of the year will be unpaired. duckling--byplumage. Eleven "pairs" remained to- Intense competition for these females could force getherfor an entire 1-h focalanimal sample, four for males to begin courting them at the duckling stage at least2-3 days,one for at least16 days, and onefor in order to obtaina mate,even though courtshipand at least22 days.Additional adult male-femaleduck- pairing with ducklingsmay be costly. ling or adult male-juvenilefemale "pairs" that could Further observationslead me to suspectthat many not be identified consistentlywere commonat all yearling malesmolt before or soonafter arriving on studysites late in the breedingseason, when almost the breeding grounds,try to establishpair bondsby no adult pairs were present. courting ducklings,and then breed in their second I believe that unpaired adult male Chiloe Wigeon year when the female has matured.Observations of directcourtship displays at ducklingsas part of a pair groupsof molting Chiloe Wigeon early in the breed- formationprocess and that some females form bonds ing seasonin southern Argentina (Weller 1975; J. with malesat an exceptionallyearly age.To explain Fjeldsapets. comm.; pets. obs.) and the unworn plum- this early courtshipand "pair" formation,it is ap- ageof courtingmales, which was often different from propriateto considerthe costsand benefitsto indi- that of male parents,support this idea.First-year males viduals(Rohwer and Anderson1988). Young females differ in their plumage from older birds in American maybenefit by being"paired" because the maleses- Wigeon (Wishart 1983) and some male EurasianWi- cortingthem droveaway othermales and were some- geon (A. penelope)delay breeding until their second timesvigilant, although less so than parentalmales. year (Cramp and Simmons1977). Data on the ages Beforeducklings accepted courting males, however, and life histories of courting male Chiloe Wigeon are theywere sometimes endangered by separationfrom needed to confirm this hypothesis. their familiesand were alsointerrupted during feed- More extensivedata on sex ratios,adult pair-bond ing or restingperiods. Adult malesseemed to incur durations, and the identities of courting males are costsby performingdisplays, by beingchased by oth- neededto testthe hypothesisthat extrememale-male er malesand parents,and by beingvigilant and de- competitionfor matesresults in courtshipand pairing fending exclusiveaccess to a duckling. Males may with ducklings.Ultimately, costsand benefitsshould benefitby pairingearly if they are more likely to be measured in terms of time, energy, and future obtaina mate,and if they expendless energy or gain reproductivesuccess for adult malesand young fe- otherbenefits as "paired" ratherthan courtingmales. males that have paired. The presenceof long-term Parentsincurred costs as they defendedthe cohesion pair bonds, highly skewed sex ratios, and delayed of the family unit, but may benefit by promoting breeding suggestedhere for Chiloe Wigeon may be "pair" formationbetween their ducklingsand high- unique to this species.Careful observationof other qualitycourting males. In manyways, the breaking little-known Southern Hemisphere waterfowl, how- of bonds between parents and their ducklings by ever,may reveal other variations on the socialsystems courtingmales was similar to theattempt of unpaired typical of their northern relatives. males to break bonds between paired adults, with I am grateful to F. McKinney, A. Pusey,L. Oring, October1991] ShortCommunications 973 and two anonymousreviewers for helpful comments haviorand pairing chronologyin wintering dab- on various drafts of the manuscript.C. Larson helped bling ducks.Wilson Bull. 95: 675-682. with the observations,and C. Bailey and D. Fookes JOHNSGARD,P.A. 1965. Handbook of waterfowl be- graciouslygranted permission to work at EstanciaSan havior. Ithaca, Cornell Univ. Press. Ramon and EstanciaPilcaneu, respectively.This re- KEAR,J. 1970. Theadaptiveradiationofparentalcare searchwas supportedby Dayton and Wilkie fellow- in waterfowl. Pp. 357-392 in Socialbehaviour in shipsfrom the Bell Museum of Natural History, Uni- birdsand mammals (J. H. Crook,Ed.). London, versity of Minnesota Graduate School, Minnesota Academic Press. Foundation-F. E. Andrews fund, Sigma Xi, and NSF LORENZ,K. 1951-1953. Comparativestudies on the grant BNS-8820065to F. McKinney. behavior of the Anatidae. Avicult. Mag. 59: 24- 28.

LITERATURE CITED MClQNNEY,F. 1985. Primary and secondaryrepro- ductive strategiesof dabbling ducks. Ornithol. AFrON,A.D., & R. D. S^¾LER.1982. Socialcourtship Monogr. 37: 68-82. and pairbondingof Common Goldeneyes,Bu- --. In press. Courtship, pair formation, and sig- cephalaclangula, wintering in Minnesota. Can. nal systemsof waterfowl. In Ecologyand man- Field-Nat. 96: 295-300. agementof breedingwaterfowl (B.D. J.Bart, Ed.). AI•TMANN,J. 1974. Observationalstudy of behav- St. Paul, Univ. Minnesota Press. iour: sampling methods.Behaviour 69: 227-267. --, & G. BREWER. 1989. Parental attendance and BELLROSE,F. C., T. G. SCOTT,A. S. HAWKINS,& J. B. brood care in four Argentine dabbling ducks. LOW. 1961. Sex ratios and age ratios in North Condor 91: 131-138. American ducks, Ill. Nat. Hist. Surv. Bull. 27: ORING,L. W., & R. D. SAYLER.In press. The mating 391-474. systemsof waterfowl. In Ecology and manage- BREWER,G. 1990. Parental care behavior of Chiloe ment of breeding waterfowl (B. D. J. Batt. Ed.). Wigeon (Anass•l•ilatrix). Ph.D. dissertation,Min- St. Paul, Univ. Minnesota Press. neapolis, Univ. Minnesota. ROHWER,F. C., & M. G. ANDERSON. 1988. Female CP,•MP, S., & K. E. L. SIMMONS. 1977. Handbook of biasedphilopatry, monogamy,and the timing of the birds of Europe, the Middle East,and North pair formation in migratory waterfowl. Curr. Or- Africa: the birds of the Western Palearctic, vol. nithol. 5: 187-221. 1. Oxford, Oxford Univ. Press. WELLER,M.W. 1968. Notes on some Argentine an- EMLEN,S. T., & L. W. OraNG. 1977. Ecology,sexual atids. Wilson Bull. 80: 189-212. selection,and the evolution of mating systems. 1975. Habitat selectionby waterfowl of Ar- Science 197: 215-223. gentine Isla Grande.Wilson Bull. 87: 83-90. GOLLOP,J. B., & W. H. M_•P,SHALL. 1954. A guide for W•SH•RT,R. 1983. The behaviouralecology of the aging duck broodsin the field. MississippiFly- American Wigeon (Anas americana)over its an- way Counc. Tech. Sect. Rep. nual cycle. Ph.D. dissertation,Winnipeg, Univ. HEINROTH,O. 1911. Beltragezur Biologie, nament- Manitoba. lich Ethologie und Psychologieder Anatiden. Proc. Int. Ornith. Congr. 5: 598-702. Received23 October1990, Accepted 27 May 1991. HEPP,G. R., & J. D. HAIR. 1983. Reproductive be-

An Interspecific Relationship Between Egg Size and Clutch Size in Birds

Tnvi M. BLACKBURN Centerfor PopulationBiology, Imperial College at SilwoodPark, Ascot, Berkshire SL5 7PY, United Kingdom

Individual birds should be selected to rear broods be compensatedby decreasedegg size. Therefore, Lack that maximizelifetime reproductivesuccess (Lack 1968, (1967) considered that such nutrient limited species Williams 1966). In many species,clutch size may be would likely show trade-offsbetween clutch sizeand limited by the number of offspringthat parentscan egg size. Smith and Fretwell (1974) also considered feed (Lack 1968, see review in Partridge 1990).How- that this trade-offwas mostlikely in speciesthat lay ever, Lack (1967) considered that precocial species large clutcheswithout parental care.Lack (1967) test- laying large clutchesshould be limited instead by ed his hypothesison wildfowl and found that species foodavailability to the laying female.Then, assuming that laid larger clutchesalso laid smaller eggs. Al- nutrient limitation, any increasein clutch size could though recently called into question(Rohwer 1988),