WHY ARE FEMALE OF PREY LARGER THAN MALES? by Dean Amadon Department of Ornithology American Museum of Natural History Central Park West at 79th Street New York, New York 10024

For everycomplex problem there is a simpleanswer and it is wrong. H. L. Mencken

The males of most birds are larger than the females,but the oppositeis true of birds of prey-the hawksand owlsof the ordersFalconiformes and Strigiformes.In thesefemales are, with few exceptions,larger than males-sometimesmarkedly so. Thesetwo ordersare only distantlyrelated, and one may assumethat this reversedsexual dimorphism evolved indepen- dently in the two orders. Further, in the predatory family Stercorariidae,the skuas and jaegers (hunters), the female is also the larger sex; this family is related to the gulls (Laridae),not to hawksor owls.• In all these raptorialbirds the clawsof the feet are modifiedinto curved,pointed talons, which are employedto seize,kill, or aid in killing the prey upon which they feed. In another somewhatpredatory family, the Laniidaeor shrikes,the bill is hookedas in raptors,but the claws are not modified into talons. The sizes of the sexes are not reversed in shrikes. There are hundreds,indeed thousands,of other speciesof birds rangingfrom flycatchers and swallowsto storkswhich are also "carnivorous,"and yet in which the male is the larger sex. But these ,with only occasionalor rare exceptions,capture inoffensiveprey- often insects,snails, and the like. Though predatorsin the literal sense,unlike the truly rapaciousbirds of prey, they do not haveto subdue,kill, and rend their prey; and they have not evolved the hooked beak and talons-or the fierce disposition that goeswith such weaponry. To analyze in detail the data supportingreversed sexual dimorphismin raptors and the possibleexplanations for it would require a volume. Here the intention is merely to sum- marize some of the hypothesesthat have been advancedand to evaluatethem briefly. Neither have ! presentedin detail the raw data which may be found in variousreference works. It may be noted, however,that there is a need for further quantitativedata on size dimorphismin raptors.This is particularlytrue of weights,which reflect the dimorphism better than do linear measurements,such as wing length. The family Stercorariidaeillustrates this point. Linear measurementsoften show superiorityin size of the female in the skuas (Catharacta)but not in the smallerjaegers (Stercorarius) (see, for example,Ridgway, 1919:681-695).When, however, adequate samples of weightsof the threespecies of jaegers were assembled(Maher, 1974:11-12),it becameevident that femalesare appreciablyheavier than males in this genusalso. Or, as anotherexample, the wing length of the male American Goshawk(Accipiter gentilis atricapillus) averages about 97 percentof that of the female;it weighsa full 20 percentless. 1In thefrigate-birds (Fregatidae) females are larger than males. Since these birds are some- what predatory,it hasbeen assumed (Amadon, 1959) that thisis anotherexample of sexual size reversalin raptorialbirds. Perhapsso, but the fact that the sameis true of the related fish-eatingboobies (Sulidae) but not of the cormorants(Phalacrocoraciidae) makes it pos- sible,if not likely, that other selectivefactors are at work in this order (Pelecaniformes). 1 RaptorResearch 9(1/2): 1-11, 1975 2 RAPTORRESEARCH Vol. 9, No. 1/2

In theiranalysis of sizedimorphism in North American owls Earhart and Johnson (1970) couchthe discussionof weightlargely in termsof the cuberoot of that dimension.This convertsweight to a statisticcomparable with linear measurements, suchas wing length. Still,it isworth emphasizing that selection for body size results from the actual weight or massof the . The authorsmentioned take thisfor grantedin their discussionof the sexualdifference in wingloading that results from dimorphism in size. Variationin Degreeof Dimorphism Raptorsvary greatly from to genusand sometimes from species to speciesin the amountof sexualsize dimorphism. Those species that are most rapacious and kill the largest preyrelative to theirown size have the greatest dimorphism. Examples range from certain ,such as the Blackand White - (Spizastur rnelanoleucus), thehorned and eagleowls (Bubo), and the larger falcons (Falco) to smallspecies of Accipiterand some small owls.In manyof thesethe maleis a full thirdsmaller than his mate, and hence called a "tiercel"by falconers.In a fewspecies ofAccipiter the male weighs only about half as much as the female. Birdsof preywith the propensityto• attacklarge and at timesdangerous prey con- commitantlydevelop the fierceand aggressive disposition necessary for suchforaging. In speciesthat takesmaller or humblerprey, such as rodents, fish, and lizards, sexual dimor- phisinis less,but oftenstill substantial. For example, in sea-eagles(Haliaeetus), various speciesof ,and othersthe weightof malesmay average one-fifth to one-fourthless thanthat of females. In primarily insect- and snail-eating raptors, or thosethat are in part scavengers,such as caracaras (Polyborus) or some kites (Milvus), size dimorphism isusually evenless and may be negligible. Thesocial habits of certainkites may also have ameliorated their aggressiveness andled to an abatementin selectionfor sexualdimorphism. The samemay be true of Eleanora's Falcon(Falco eleanorae), which, as Dr. NoelSnyder (in litt.)has pointed out to me,shows remarkablylittle dimorphismfor a bird-eatingspecies. To besure, the fact that most of the birdsit catchesare small may be a factor.Social habits and extensive use of insectsas food mayexplain the fact that in theBurrowing Owl (Athene cunicularia) males average slightly larger than females. Anotherexception is theSecretary-bird (Sagittarius serpentarius); it occasionally kills large,venomous snakes, but its claws are blunt, not modified into talons. Though classified withthe birds of prey,it maybe unrelated to them.Among vultures, which are for the most partentirely carrion eaters, the sexes are of aboutthe same size, or the male may even be a littlelarger. The claws of vulturesare long, coarse, and blunt-no longer talons. The Old Worldvultures, in particular, are closely related to otherbirds of preyand presumably derivedfrom them. It isreasonable to assume that as they gradually lost their predatory habits,they also lost their reversed sexual size dimorphism. Thecorrelation just traced is sometimes evident among the species ofa single genus. The PeregrineFalcon (Falco peregrinus) preys on birds, including large ones; it haspronounced dimorphism.More or Jessinsectivorous species of thegenus, such as the kestrels, are less dimorphic.Bird-eating species ofAccipiter are highly dimorphic; theamphibian-eating Grey Frog-hawk(A. soloensis)of the Orient is lessso. Earhart and Johnson (1970:260) found thatwithin a singlespecies, theScreech Owl (Otus asio), the chiefly insectivorous subspecies areless dimorphic than ones that take many birds and small mammals. The same may be trueof the American Kestrel (Falco sparverius) (Storer, 1966:433-434). Thus,one finds on the whole a remarkablyclose correlation in birds of preybetween degreeof rapaciousnessanddegree of reversedsexual dimorphism. Spring-Summer 1975 Amadon - Size in Birds of Prey

Why Reversed Dirnorphism ? Hypothesesfall into two main categories.The first relatesreversed sexual dimorphism to one or anotheraspect of reproductivebehavior. I continueto favor one suchhypothesis. But before discussingbehavioral explanations, I shall considerand attempt to disposeof the prevailingalternative hypothesis, namely, that dimorphismin raptorialbirds is an adaptation to widen the food niche. That is, if the largerfemale takes largerprey on the average,and the smallermale takes smallerprey, betweenthem they will utilize a greaterdiversity of the potentially available food. Such expansionof the food nichemight be important duringthe nestingseason, when the pair of raptorsis restrictedto the generalvicinity of the nestand when they have voraciousyoung to feed. It might alsobe of importanceto somespecies during winter or other periodsof possibleshortage of food. At presentthose who seein the size dimorphismof raptorsan adaptationto expandthe food niche are in the majority. This may be attributed in part to the currentpopularity of studiesembraced by suchterms as "time and energybudgets," "(avian)energetics," "adap- tive strategies,"and "physiologicalecology." There are certainly birds and mammals in which the sexesdiffer in size but eat the same thing (e.g., grazers,whether geeseor wild cattle). Among predators,however, as would be expected, large individualstend to kill larger prey than do smaller ones. Large female accipiters take on averagelarger prey than do the smaller males (Storer, 1966; Brosset, 1973), and the same is true of Circus (Schipper, 1973) and doubtlessof other highly dimorphicspecies, such as the largerfalcons. But the fact that a largehawk (or person)eats more and preferslarger bites than a smaller one is not necessarilythe reasonit is larger. Reasonsfor doubting the ecologicalor niche-expansionexplanation of reversedsexual sizedimorphism among raptorial birds are the following: First, as already summarized,the degreeof sexual dimorphismin raptorsis correlated with the relative size of the prey taken, not with the potential of the food supply for supportingdimorphism. Species that regularly or not infrequently attack and kill birds or mammalsas largeor largerthan themselveshave the greatestdimorphism. Such species range in size from sparrow-hawksto big eagles.Species that take humbler prey, even when there would seem to be as great or greater potential for broadeningthe food niche, are less dimorphic. Second,the ecologicalhypothesis fails to explain, at least convincingly,why it is the female that is the larger sex. Some (e.g., Selander,1966:139) recognizethis failure and attribute the initial impetustowards reversal to behavioralfactors, but then assumethat the dimorphism,because of its extent, must be under ecologicalcontrol. Othershave suggested (Prof. E. Mayr, pers.comm.) that sincethe male doesmost or all of the huntingin the first half or three-fourthsof the nestingcycle, it is advantageousfor it to be the smallersex because(a) there are more speciesand individualsof small than of large prey (Elton's "pyramidof numbers"),and so the male can get more food at the time he is doingall the hunting;and (b) smallerprey is more suitedto the chickswhen they are small. There are a numberof caveatsto be enteredhere. Will a largernumber of smallprey alwaysadd up to a largerbiomass, especially when muchof the prey hasto be transported to the nest?Some buteosfeed on insectswhen away from the nestbut carry only verte- bratesto the nest; insectsare too smallto make it worthwhile.Certain species commence nestingwhen the country is stillcold or evensnowbound, when the preyanimals have not yet broughtforth their young,and when smallerrodents are lessavailable. The stronglydi- morphicGreat Horned Owl (Bubo virginianus)in the northernparts of its rangebegins nestingin late winter when the ground is still coveredwith snow. Not until weeks later, whenthe youngare well grownand when the femaleis alsohunting, are young prey 4 RAPTOR RESEARCH Vol. 9, No. 1/2 available.Nor is smallprey necessarilymore suitedto the chicks.The female might find it easierto tear bits of fleshfor the youngfrom, say,a rabbit, than from a mouse. Dr. Ned Johnson(pets. comm.)has suggested to me that followingan initial selectionfor largersize in the femalefor behavioralreasons, females that pairedwith smaller,more agile malesthat were better foragerswould leavethe mostprogeny. Skill in aerial food passingin courtshipmight be part of suchselection. This hypothesisis attractivefor suchspecies as harriers(Circus) or somefalcons, but if one considersit in the gamut from smallowls to eaglesit is lessconvincing. A raptor that is a little lessagile may offsetit by catchingbigger prey. The largerfemale doessome effective hunting after the youngare well grown,as well ashunting for herselfduring the nonbreedingseason. Reynolds(1972) concludedthat if dimorphismis an adaptationto broadenthe food niche, then it will be greatestin bird-catchingspecies, such as the Sharp-shinnedHawk (Accipiter striatus),in which the sizeof the predatorapproximates that of the agileprey. But surelythe vastmajority of the birds caughtby this speciesand by relatedones, such as the EuropeanSparrow-hawk (A. nisus),weigh no more than a third or perhapshalf asmuch as the hawk. The argumentis perhapsundercut to someextent by Reynolds'sevidence that the largerfemale, when she doescommence hunting later in the cycle,is effectivebecause her size enablesher to take a greater size range of prey. And as with other ecological hypotheses,the pronounceddimorphism in many speciesthat catch few birds remains unexplained.Weight data (Glutz, Bauer,and Bezzel,1971:173,642) indicate,for example, that in the Europeanrace of the GoldenEagle (Aquila chrysaetos)and evenin the White- tailed Sea-eagle(Haliaeetus albicilla) the averageweight of the malesis a full 25 percentless than that of the females. Third, a surveyof the kingdom suggeststhat sexualdimorphism in sizeis usually correlatedwith mating behavior(Areadon, 1959). In general,males are more aggressivethan females;they competefor mates. Largerand strongermales have an advantagein perpetuat- ing their genes.This distinctionmay be lessapparent in birds than in mammals,but it exists. Even in the occasionalinsect in which males contest physically for mates, selection has producedlarger size in that sex (Beebe, 1947). The increaseddimorphism often found in polygynousor promiscuousspecies, such as some of the largergrouse (Tetraonidae) or pheasants(Phasianidae) or bustards(Otididae), further illustratesthe point. In highlypoly- gynousseals (Phocidae), such as the elephant-seals(Mirounga) or northernfur seal(Callo- rhinus), two generathat are not closelyrelated, males weigh several times as much as the females.In thesepolygynous birds and mammals,males are to a degreeexpendable; only a few are needed for fertilization. Hence, sexual dimorphism can proceed further than in monogamousspecies. In the latter, amongseals as well asbirds, dimorphismis much less.An ecologistwho saw in the enormousdimorphism of someof the polygynousseals an adapta- tion to broadenthe food nichewould surelybe mistaken! When the parental-carerole is reversed,as in phalaropes(Phalaropidae) or button-quail (Turnicidae),and the malesincubate and carefor the chickswhile the femalescompete for mates,it is the femalesthat are the larger,and, when there is a distinction,the more brightly colored sex. Female button-quail are so pugnaciousthat they are pitted like miniature gamecocks.The reversedsexual dimorphism of raptors is not associatedwith reversedroles in parental care. The instancesabove are merely cited to further illustrate the responsiveness of size to selectionby reproductivebehavior. The skuasand jaegers(Stercorariidae) seem to confirm this observation.In the related gulls (Laridae) males are larger than females.Yet, when the skuasbecame even more predatory than gulls, and, especially,one supposes,when they developedsmall, curved "talons" that could pose a threat to the female in pairing and territorial encounters,selec- tion came to favorlarger size in the female.Perdeck (1960), who studiednesting behavior in Spring-Summer1975 Amadon - Size in Birdsof Prey

the GreatSkua (Catharacta skua), is quotedon thispoint below. Thoughprobably without immediaterelevance to the situationin birdsof prey, onemay cite Jehl's(1970) study of dimorphismin certain Scolopacidae.In the Stilt Sandpiper (Micropalamahimantopus) and the LeastSandpiper (Erolia minutilla) he foundthat individ- ually smallmales and largefemales pair first and aremost successful in nesting.Again, if he is correct, the dimorphismhas its basisin reproductivebehavior, not in ecology.The fact that the femaleis the largersex in theseshorebirds is unrelatedto the situationin raptors, but reflectsthe trend towardsreversal of parental care in shorebirdsthat has becomecom- plete in suchgroups as the phalaropes. Thus there is variedevidence that sexualsize dimorphismamong higher animalsis commonlythe resultof patternsof matingbehavior. The burdenof proof is upon thosewho suggestthat the sameis not true of raptorial birds. Fourth,there are doubts as to the theoreticalbasis of the ecologicalor niche-expansion theory.One assumesthat intraspecificcompetition among birds is regulatedby suchadapta- tions as territory, peck order, delayedreproduction, cannibalism, and the like. Such com- petitionis often densitydependent and will not lead to extinction.Interspecific competi- tion, on the otherhand, may posea directthreat to survival,and thousands of specieshave doubtlesssuccumbed to it sometimes,to be sure, when also beset by environmental changes.Because the environment does change seasonally, annually, and over longer periods, successfulspecies evolve various safety factors. One may be a body sizethat is not rigidly limited by food or environment.Such ecological amplitude may be especiallynecessary in top-levelpredators whose food supplyis often moreuncertain and fluctuatingthan that of the grasseaters which sustainthem. Hencein speciesin whichmating behavior makes sexual dimorphism in sizeadvantageous (andthis includes most birds and mammals), the ecologicalniche, one supposes, is often or perhapsusually molded by selectionto accommodatethe size variation and not vice versa. Suchmodifications may entail increased size of territory,heightened interspecific aggression, etc. Somespecies, to be sure,adjust with difficulty;perhaps this is why certainraptors, such asthe Ayres'Hawk-eagle (Hieraaetus dubius) or the TaitaFalcon (Falco fasciinucha), are so rare. Others become extinct. To statethis the otherway around,would not interspecificcompetition be leastwhen a speciesis clusteredaround its optimalcompetitive size, without sexual-size dimorphism? If so, the widespreadoccurrence of dimorphismamong higher animals may be regardedas a behavioralnecessity. To what extenttheoreticians have dealt with suchquestions I do not know. Schoener'scontributions (e.g., 1968, 1969a, 1969b) seemto dealwith rathercom- plex situations,and his equations have so many variables as to renderconclusions from them tentative. Evenif interspecificcompetition isdisregarded, sexual differentiation in food or foraging nichemay not be advantageousin the longrun. If the food supplychanges, there are two thingsthat can go wrong, not one.It maybe moredifficult for sucha dimorphicspecies to adjustto changingfood resources or levelsof interspecificcompetition. Thuswhen it is demonstratedthat the smaller male, as might be expected, requires fewer caloriesthan does the female (Mosher and Mattray, 1973) of theBroad-winged Hawk (Buteo platypterus),one doesnot needto acceptthe suggestionthat the smallersize of the male reflectsdirect selectionto lessencaloric requirements. Why would suchselection have affectedthe maleonly? Why do speciesbecome larger in thefirst place if caloricrequire- mentsare of suchcrucial importance? This, of course,is not to denythe importance of such studiesin determiningphysiological parameters. 6 RAPTOR RESEARCH Vol. 9, No. 1/2

The fossilrecord contains many examples(and in diversegroups!) of a trend towards increasingsize. Larger individuals may securemore mates, or they maybe ableto shoulder companionsaway from food or othernecessities (Simpson, 1944:86). The trendtowards largersize is a generalone, and the bluewhale (Balaenoptera musculus), a planktoneater, has become the largestknown animal of all time. In thousandsof diverseanimals, and perhapsone mightinclude plants, size has not beenso rigidlycontrolled by foodor other environmentalfactors as to precludegradual change. Often it has continuedto changeover vastperiods of time. To be sure, a number of examplesof sexualdimorphism in birds apparentlyrepresent adaptationsto expandthe availablefood resources(summarized by Selander,1966). These usuallyrelate to the primary food-gatheringorgan, the bill, rather than to generalsize, and the speciesconcerned are chiefly inhabitantsof depauperateislands where interspecific competitionwill usuallybe lessthan on continents.Even so, the evidencethat the observed sexualdifferences in foragingare the causeof the dimorphismis circumstantial.Perhaps the species-poorhabitats merely providemore scopefor divergenceresulting from socialor sexual behavior (dominance,peck order, mating systems,facilitation of sex recognition, etc.). ! do not, of course, wish to deny that food availability or lack of availability may sometimeshave had somethingto do with degreeof dimorphismin raptorialbirds. Earhart and Johnson(1970) found some racesof Bubo virginianusto be slightlymore dimorphic than others; this may reflect some environmentalinfluence, but it could be a result of mere chance.

Behavioral Explanations It will be clear from what has preceded that I continue (Brown and Amadon, 1968:26-28) to think that the explanationfor the reversedsexual size dimorphismof rap~ torial birds is to be soughtin the relationshipsof the male and femaleat the time of pair formationand perhapsthroughout the reproductiveperiod. Before discussing this, we may disposeof hypothesesbased on other aspectsof behavior. First, it has been suggestedthat male birds of prey, lacking "maternal" instinctsand accustomedto killing small animals,may pose a threat to the life of nestlings.Selection might then favor larger size in the female so that she can preventany suchattempts by the male. Perhapssuch reflectionswere influencedby knowledgeof certainmammals, such as cats and bears, in which the males are, in fact, a threat to the young. There is now ample evidencethat evenin raptorswith pronouncedsexual dimorphism the male often incubates to some extent and may even have a brood patch. He sometimesfeeds the chicks and certainly posesno threat to them. A secondhypothesis suggests that selectionhas favoredlarger size in the female raptor, the better to enable her to protect the nest and its contents from predatorsof other species. Thissuggestion is moredifficult to disprove,but I donot thinkit will passmuster. It istrue that the femalespends more time at the nestthan the male,but why, if nestdefense explains the larger size of female raptors, should the dimorphismbe greatestin fierce speciesthat catch relatively large prey? It might be arguedthat becausethey are fiercer, they defend the nest more vigorously;therefore selection for increasedsize is more severe.This is rather tenuousreasoning. To be sure, two fierce species,the Great Horned Owl and the Goshawk, are the North Americanraptors that mostoften attackhumans climbing to the nest.But it is also true that many raptorswith rather pronounceddimorphism, such as the two speciesof North American eagles,are lackadaisicalabout nest defense,at least insofar as humansare concerned. Spring-Summer 1975 Amadon - Size in Birds of Prey

Scattered evidence indicates that males do participate in nest defense. In Great Horned Owls both members of a pair sometimesattack an intruder, striking alternate blows. Both membersof a pair of FerruginousHawks (Buteo regalis)attacked a coyote that was un- suspectinglyapproaching their hillsidenest (Angell, 1969). A male Merlin (Falco colum- barius)sometimes appears, screaming in protest,when an intruder is still far from the nest. When raptors nest or attempt to nest in captivity, both membersof the pair may become very aggressivetowards man or any intruder, aswas true of an observedpair of Harpy Eagles (Harpiaharpy]a) (Dr. WilliamG. Conway,pets. comm.). Nest defenseis a two-edgedsword: better to flee and live to nestanother day than to be killed by a nest predator. Golden Eaglesoccasionally bring Great Horned Owls to their youngas food; perhapsthese were owls that defendedtheir own neststoo vigorously!The speciesof raptors found in any particular area usually exhibit wide divergencein di- morphism;it seemsunlikely that nest defenseis more important to some of them than to others. Thirdly, the suggestionis madethat contestsbetween male raptors are primarily brilliant aerialmaneuvers and that smallersize in the male, far from beinga handicap,is an asset.This is an attractiveidea asapplied to a dashingfalcon, but not for a big eagleor a hornedowl. Further, Cade (1960) found that large falconsdo in fact make bodily contact in aerial encountersand that sheerphysical prowess is the decidingfactor. We turn, finally, to the hypothesis that reversedsexual dimorphism is the result of reproductiveinteractions between the sexes.Pair formation may prove to be the decisive phase. In highlydimorphic birds as regards color-many pheasants, for example-sex_recognition may be innate;but in the numerousspecies in whichthe sexesare externally alike, and this includesmost raptors,initial pair formation is often-as the late R. C. Murphy onceput it-a matter of trial and error. Typically, the male takes up a territory and, by displaysand vocalizations,warns off other malesand attracts females.He is hostile to any approaching conspecific,treating it as a potential rival and buffeting it or otherwiseseeking to chaseit away. If the approachingbird is a femaleprepared to mate, shebehaves submissively, refuses to be repelled,or thereafterreturns and a pair is formed (Lack, 1940). If this is the sequencein raptors,and many more observationsare needed,may not the femalebe in actualperil from a male armedwith talonsand (judging from birdsin general) innately more aggressivethan she?Indeed, the male might first appraisean approachingbird of his own or smallersize as potential prey! In suchspecies selection would, one assumes, have favoredfemales that were aslarge as, or eventuallylarger than, the male so that despite their somewhatmore submissiveor passivenature at the time of pairing, they would not be intimidated or imperiledby the male. This would be especiallytrue of speciesthat unhesitat- ingly attackbirds as large or largerthan themselves. It is in suchspecies, as has been already emphasized,that the superiorsize of the femaleis usuallymost pronounced. If such is the case,one might ask why sexualdimorphism is substantialin somebuteos, fish eagles,and the like, which subsiston rather humble prey. Theseraptors are, however,all armedwith dangeroustalons, and most of them do at timesattack largebirds or mammals. A few exampleswill be given.A Red-tailedHawk (Buteojamaicensis) when flushed,flew off carryingthe body of one of its own specieswhich it had apparentlykilled and was eating (Clevengerand Roest,1974). A WhiteHawk (Leucopternis albicollis), a speciesthat normal- ly feedson smallsnakes and the like, killed a tinamou(Tinamus major), a bird that would weighmore than the hawk (Lamm, 1974). The AmericanKestrel occasionally tackles a bird as heavy as itself. The (Haliaeetusleucocephalus), though partial to fish, not 8 RAPTOR RESEARCH Vol. 9, No. 1/2 infrequentlypursues and kills largebirds; some individual eagles do so habitually.Further, Accipiter and Falco, instructive though they may be, do not provide all the necessary evidence.For example,speculation about the correlationof increasedsexual dimorphism in colorand reduceddimorphism in sizein Falco sparveriuscannot ignore other genera(or even other speciesof kestrels)in whichno suchcolor dimorphism exists. Careful observationsof both wild and captiveraptors are needed.The behaviorof captive birds may be significantbut must be interpretedwith caution. In the wild, for example, courtshipor pairingflights may covermiles of terrain.Willoughby and Cade(1964) in an attempt to study dominancein relation to size dimorphism,paired a male of the northern race of Falco sparveriuswith the female of the smallersouthern subspecies, thereby forming a pair of about the samesize. Behavior did not seemabnormal, but thisis a speciesof rather moderatedimorphism. It might be arguedthat thingshave swungtoo far. In courtshipfeeding in the Peregrine Falcon the male sometimesappears terrified of the largerfemale; she snatches the prey from his graspas he speedsby. Where attempts have been made to breed variousspecies of Accipiter in captivity, the male, lackingsufficient room to flee, is sometimeskilled by the female. Nonetheless,under natural conditions,one assumesthat the existing degreeof dimorphismin each speciesis the optimum under existingselection pressures. As noted earlier, Perdeck(1960:129) in his study of Catharactaskua cameto the con- clusion set forth above as to the reasonfor reversedsexual dimorphism. If he was correct, the samecould be true of hawksand owls.He summarizedas follows (slightly paraphrased):

Perhapsthe piraticaland sometimespredatory habits of the skuahave something to do with the reversedmale-female size ratio, as comparedwith the gulls.A predatory bird needsa high aggressivenesstowards other birds that may not only be its equalin size but also be of the samegeneral appearance (skuas prey upon gulls). But this extreme aggressivenessis a drawbackin pair formation and mating. A male has to perform the most 'aggressive'part of the mating. For the male to be in generalmore aggressivethan the femalehas, therefore, survival value. But he must, of course,not be too aggressive!The chanceof this occurringis greaterin a speciesthat, becauseof its feedinghabits, requires an extra 'dose'of aggressiveness.For sucha speciesa reduction in the sizeof the male relative to that of the female may have survivalvalue.

In passing,it may be noted that sex recognitionin predatorymammals is basedon other cues,probably initially olfactoryones, and thereis no reverseddimorphism. To attribute such a fundamentalcharacter as sexualdimorphism in sizein raptorsto a selectivepressure that operatesonly duringthe relativelybrief periodof pair formationand matingmay seemunrealistic. Yet manybirds (and mammals)are encumberedthroughout adult life with "ornaments"used only at that season(for examplepeacocks [Pavo] ). At the sametime it is possiblethat the reverseddimorphism of raptorsmay enhancereproductive successthroughout the entirenesting cycle. Cade (1960) wrote asfollows:

A reproductivelysuccessful pair bond can result only when the femalefalcon is clearly dominantto the male... and when the male makesa biologicallyadequate adjustment to his subordinaterole in the pairing situation.... Female dominance appearsto function in some as yet unknownway in holdingthe male to his role as food providerfor the female and the young.... On the average,females that mate with males smallerthan themselveswere able [one supposes]to maintain better dominancerelations, and more progenyof suchpairs survived .... Even in Kestrels,in Spring-Summer 1975 Amadon - Size in Birds of Prey

which the dimorphismis less, the fern:des"assert a vigorousdominance over their mates."

In the Goshawk,Schnell (1958:385) found that the female had a specialcry with which she "dismissed"her mate; and shedid not begin feedingthe chicksuntil after he had left the vicinity of the nest. Kemp and Kemp (1975) observeda female Ovampo Sparrow-hawk (Accipiter ovampensis)threaten her mate when he remained at the nest after she had finishedeating a prey item the male had brought. The male then left immediately.But there are hundredsof speciesof birds in which one sex, usually the female, assumesthe entire duties of parental care without stimulus or assistancefrom the mate. There are a few observationsof male that continued to feed the chicksafter the death of the female, and there is no real evidencethat physical(or sexual) dominanceby the larger female has anythingto do with the fact that her mate hunts for her and for the chicks. How may one demonstrateconclusively whether behavioralor ecologicalfactors or both led to reversedsexual dimorphism in raptorial birds? Comparison of the degree of di- morphismwith the opportunitiesfor ecologicalexpansion might provide clues, e.g. com- parisonof the dimorphismand food resourcesof the subspeciesofButeo jamaicensison the tiny islandof Socorrooff the Mexicancoast with thoseof one of the mainlandsubspecies. Mating and parental behaviormay be observedboth in the wild and amongthe raptorsthat are being bred with greater successin captivity. The data acquiredmay prove difficult to interpret at leastinitially, but will alsobe valuablefor other purposes. I am grateful to Messrs.John Bull and Mark Fuller and the otherswho helped me with this paper.

Summary In most birds of prey, includingsuch unrelated ones as hawks (Falconiformes),owls (Strigiformes),and skuas(Stercorariidae, Charadriiformes), females are larger than males, sometimessubstantially so. This reversedsexual size dimorphism in raptors is greatestin speciesthat pursue,kill, and rend large, active prey; lessin those that kill small rodents or insects;and absentor virtually so in vultures.In somespecies it hasbeen shown,as would be expected, that the larger females catch on the averagelarger prey than the smallermales. Some ecologistsconclude from this that the dimorphismis an adaptationto increasethe total size range of prey availableto a raptorial species.This hypothesisdoes not explain why the dimorphism is greatestin fierce speciesthat kill large birds and mammals,why in some dimorphic speciesthe sexesapparently take food of the same size, or why it is the female that is the larger sex, althoughan attempt has been made to correlatethe last point men- tioned with ecology. Furthermore,in highervertebrates in general,sexual dimorphism in sizeis correlatedwith matingbehavior, not with foragingecology; it is usuallya resultof competitionby malesfor mates.When it is the femalesthat competefor mates, then that sex is the larger.Although femaleraptors do not competefor mates,they are largerthan the males;the explanationis probably still to be found in mating behavior.Raptors have becomeunusually aggressive birds, presumablybecause they hunt and kill activeprey at somerisk of injury to them- selves.Furthermore the male, as in birds in general,may be expectedto be the more aggressivesex, at least at the time of mating and pair formation. Armed as he is with formidablebeak and talons,he would posea threat to the physicalwell-being of the female duringpair formation,which in birdsinvolves aggression by the maleduring its earlierstages. To offset this, selectionwould favor greaterphysical prowess in the female,and hence that sexhas become the largerone, especiallyin the most aggressivespecies. 10 RAPTOR RESEARCH Vol. 9, No. 1/2

References Cited

Amadon, D. 1959. The significanceof sexualdifferences in size amongbirds. Proc. Amer. Phil. Soc. 103:531-536. Angell, T. 1969. A study of the FerruginousHawk. LivingBird 11:225-241. Beebe, W. 1947. Notes on the HerculesBeetle. Zoologica 32:109-116. Blotzheim, U.N.G. yon, K. M. Bauer, and E. Bezzel. 1971. Handbuch VogelMitteleuropas. Falconiformes,vol. 4. Frankfurt: AkademischeVerlag, 943 pp. Brosset,A. 1973. Evolutiondes Accipiter forestibres de l'est du Gabon.Alauda 41 '185-202. Brown, L.H., and D. Amadon. 1968. Hawks, eaglesand falcons of the worM. New York: McGraw-Hill. Cade,T. J. 1960. Ecologyof the Peregrineand Gyrfalconpopulations in Alaska.Univ. Calif. Publ. Zool. 63:151-290. Clevenger,G. A., and A. I. Roest. 1974. Cannibalismin Red-tailedHawk. Auk 91:639. Earhart, C. M., and N. K. Johnson.1970. Sex dimorphismand food habitsof North Ameri- can owls. Condor 72:251-264. Jehl, J. R. Jr. 1970. Sexual selectionfor size differencesin two speciesof sandpipers. Evolution 24:311-319. Kemp, A. C., and M. I. Kemp. 1975. Observationson the breedingbiology of the Ovampo Sparrow-hawk,Accipiter ovampensis Gurney (Aves: ).Ann. TransvaalMus. 29:185-190. Lack, D. 1940. Pair-formation in birds. Condor 42:269-286. Lamm, D. W. 1974. preyingon the Great Tinamou.Auk 91:845-846. Maher, W. J. 1974. Ecologyof...jaegers in northernAlaska. Pac. CoastAvifauna No. 37, 148 pp. Mosher,J. A., and P. F. Matray. 1974. Size dimorphism:A factor in energysaving for Broad-wingedHawks. Auk 91:325-341. Mueller, H. C., and D. D. Berger. 1970. Prey preferencesin the Sharp-shinnedHawk; the rolesof sex, experienceand motivation.Auk 87:452457. Perdeck,A. C. 1960. Observationson the reproductivebehaviour of the Great Skua or Bonxie,Stercorarius skua skua (Briinn.), in Shetland.Ardea 48:111-136. Reynolds,R. T. 1972. Sexualdimorphism in accipiterhawks: A new hypothesis.Condor 74:191-197. Ridgway,R. 1919.The birdsof North andMiddle America, Part 8. U.S. Natl. Mus.,Bull. 50. Schipper,W. J. A. 1973. A comparisonof prey selectionin sympatricharriers, Circus, in western Europe.Le Gerfaut 63:17-120. Schnell, J. H. 1958. Nestingbehavior and food habits of Goshawksin the SierraNevada of California. Condor 60:377-403. Schoener,T. W. 1968. Sizesof feedingterritories among birds. Ecology 49(1): 123-141. 1969a. Models of optimal size for solitary predators.Amer. Natur. 103(931): 277-313. 1969b. Optimal size and specializationin constantand fluctuatingenviron- ments. In: BrookhavenSymp. in Biol., Diversity,and Stability in EcologicalSystems. Upton, N.Y.: BrookhavenNational Lab. pp. 103-114. Selander,R. K. 1966. Sexualdimorphism and differentialniche utilization in birds.Condor 68:113-151. Simpson,C. G. 1944. Tempoand mode in evolution.New York: ColumbiaUniversity Press. Spring-Summer 1975 Schaughency - Hawk Kills Egret 11

Storer, R. W. 1954. Weight,wing area, and skeletalproportions in three accipiters.Proc. Int. Ornithol. Congr. 11:287-290. 1966. Sexual dimorphismand food habits in three North American accipiters. Auk 83:423-436. Willoughby,E. J., and T. J. Cade.1964. Breedingbehavior of the AmericanKestrel (sparrow hawk). Living Bird 3:75-96.

HARRIS' HAWK KILLS EGRET by C. B. Schaughency RR 2, Box D-1 Delwood Road Chester, New Jersey 07930

The Harris' Hawk (Parabuteounicintus), though it capturesmany small animalsand is even said to take carrion, at times tackles larger prey without hesitation, both in the wild and when trained for hunting, as the following experienceillustrates. On the morning of January 25, 1975, I was photographingin a swampyarea along a road in San Luis Potosi, Mexico. A Harris'Hawk flew in, swoopedon a lone nearbySnowy Egret (Egretta thula), and carriedthe squawkingheron about one hundred yards to a low stub.When I approached,the hawk carried its prey about fifty yards further, landingin a palm tree and droppingthe egret, still alive, at its base.Six other snowiesflew around the victim excitedly, and then a CommonEgret (Casmerodius albus) flew in andlit nearbyfor a coupleof minutes.After ten minutesthe hawk droppeddown, seizedthe egretwhich was still able to squawkonce or twice more, and laboriouslycarried it off behind a screenof trees,where the hawk ap- parentlyfed. A quarterof an hour later the hawk, unencumbered,returned to its perchin the palm. Probablythe egretweighed about half asmuch as the hawk.

RaptorResearch 9(1/2): 11, 1975