IDENTIFICATION

ARADOXICALLY,THESMALLER. more The identification of slowlymoving Sharp-shinned Hawk (,4. striatus)beats its wingsmore rapidly North American Accipiters than the rapidly moving Cooper's or Goshawk (A. gentilis). The reasonsfor this are complex.During the downstroke in flapping flight, the tip of the wing Seventyman-years of watchingand trappingdistilled herewith: movesmore rapidly throughthe air than how to, and how not to, identify Accipiters the body.The larger the bird. the longer the wing and the greater this differential. A large bird can thus moveits "propel- ler" rapidly with a relatively slow wing Helmut C. Mueller, Daniel D. Berger and GeorgeAllez beat. At a givenrate of flapping, a larger wing is subject to considerablygreater mechanicalstresses; an extremelylarge wing flapped very rapidly would disin- N NUMEROUSOCCASIONS wehave ences in plumage) is size. As a bird tegrate, unless it were constructedvery been asked to identify the three increasesin size,wing area increasesas a sturdily. and thus very heavily. prohibit- speciesin the genus Accipiter (most functionof the squareof a linear dimen- ing flight. Rapid flapping of a verylarge recentlyby the editor of Arnerican Birds, sion, and volumeand weightincrease as wing would also require prohibitively hence this article). Our usual answer to a function of the cube of the linear largeflight muscles. this questionis: "... spendmany hours dimension. If general body form and A Sharp-shinnedHawk in flight is in the field with an expert and carefully shaperemain constant, doubling a bird's buoyant.owing to its light wing loading; watch a few thousand correctlyidentified length resultsin a wing area four times and flaps rapidly owing to its short birds." We have found, unfortunately. asgreat and a bodyweight eight times as wings. At the other extreme. Goshawk that not all "experts" are really expert. heavy. Wing loading (weight per unit flight is more direct and "purposeful" The few expertswe can personallycertify wing area) is doubled.The heavierthe becauseof the heavy wing loading. The are all trappers and bandersof hawks. wing loadingthe faster the bird can and Goshawk flaps slowly for it has long Banding stations for the capture of must fly. This more rapid flight is not wings.Cooper's Hawks possessinterme- migrating hawks do not have room to obvious to the human eye because a diate flight characteristics.These differ- entertain bird watchers and the smaller object movingat the same speed encesin flight characteristicsare appli- researchersare too busy to be instruc- as a larger one is perceivedas moving cable for all birds of similar configura- tors. Seriousstudents of hawk migration more rapidly. The top flight speedsof tions but different sizes. For example. usually concentrate their efforts at one hummingbirdsand Starlingsare about the Ring-billed Gull can be readily dis- localityand rarely visit other stations;we equal, yet the hummingbirdappears to tinguishedfrom the larger Herring Gull thus rarely have the opportunityto eval- movemuch more rapidly (Hertel, quoted on the basisof flight characteristics,with uate the abilities of other workers.Trap- in Welty, 1975). a little practice. pers and bandershave the advantageof verifying some of their identifications with the bird in the hand. In our combined, morethan 70 man-yearsof watchingand trapping hawks, we have banded more than 6000 Accipiters and observedmany thousandsmore. During migration we trap more than 20 per centof the Accip- iters we observe.Our expertisewas not acquired quickly and easily. The first Cooper's Hawk (Accipiter cooperil) we caught, about 30 years ago, became a Broad-wingedHawk (Buteoplatypterus) after an embarrassinglylong examina- tion in the hand. The characteristicswe use to identify Accipiters are all related to size. The three Accipiter species differ little in body form or shape,and relativesize of wing and tail. Thus the only easily noticeable difference between the three species(ignoring, for now, a few differ- Sharp-shinnedHawk. Fire Island, N.Y. Photo/Harry N. Darrow.

236 AmericanBirds, May 1979 banding. We routinely measureweight. tail length. and wing chord. The latter measurement is the distance from the bend of the folded wing to the tip of the primaries. To arrive at "estimated body length" we havetaken the averagelength for a speciesgiven in Peterson(1947) and then used the ratio of this measurement to our thousands of measurements of wing chord to arrive at the statistics presented in Fig. 1. The body length presented in Peterson is that of a museumspecimen. which is roughlythe distancefrom bill tip to tail tip of a bird lying on its back. We have used this, rather than other published measure- ments, becauseit most closely approx- imates the length of a bird in flight, whichis hog'mostAccipiters are observed. We have also indicatedthe lengthsof Adult Cooper's Hawk, Oct., 1974. Photo/Fred Ti!!y. some common "roadside" birds in Fig. ETHINK THAT MANY birdwatchers two standard deviations which includes 1. From this we can seethat a d Sharp- wouldbe able to identify the three 95.4% of the birds, and the line extend- shinnedHawk is slightly longer than an speciesof Accipiter by watching a leg ing to either side indicates three stan- American Robin (Turdus migratorius) •hundredcorrectly identified Sharp- dard deviations. which includes 99.7%. but shorterthan a Mourning Dove (Zen- shinneds,Cooper's and Goshawksand or virtually all birds. The purposeof this aida macroura). A 9 Sharp-shinned by paying close attention to size and method of presentationof data is to show Hawk is about as long as a Belted King- flight behavior.Unfortunately there are the variationbetween sexes of a species fisher (Meœacerylealcyon) and shorter complications. Female Accipiters are aswell as the differencebetween species. than a Common Flicker (Colaptesaura- considerablylarger than males; indeed, For example. all Cooper's are larger tus). (Remember that kingfishers and the differences between sexes within a than all Sharp-shinnedsand most d flickers have longer beaks than hawks, speciesare almostas great as the differ- Cooper'sare considerablylarger than and this is included in the measure- encesbetween species. If onecan identify most 9 Sharp-shinneds.Further, all 9 ment.) Cooper's Hawks average shorter Accipiters by size, one can also deter- Sharp-shinnedsare larger than all d than CommonCrows (Corvus brachyrhy- mine sex. Sharp-shinneds.All three species are chos), males considerablyso. Goshawks Figure I shows the estimated body clearly separableby size. Although the are considerablylonger than crows. lengthof malesand femalesof the three sexesoverlap in estimatedlength in the Accipiterspecies in diagrammaticform. Cooper's and Goshawk, note that fewer HEREIS NOSINGLE reliable good For each sex of each species,the long than 5% of the birds fall into the zone of measurement of size in birds: we verticalline indicatesthe mean (average) overlap. We might remind the reader feel that weight is one of the better ones. length. The broad box to either side of that Fig. I illustrates "estimated" Figure2 presentsthe weightsof the three. the mean indicates one standard devia- length.The lengthof a bird is difficult to speciesof Accipiters. This figure is tion, which (statistically)includes 68.3% measurein a live or dead specimen.and drawn from data gathered at Cedar of all the birds of that speciesand sex. we have not attemptedto measurebody Grove Ornithological Station. Wiscon- The narrow box to either side indicates length on birds we have capturedfor sin. Averageweights of non-raptorial Cooper'sHawk9 Goshawk9 Sharp-shinned Hawk

d 9

Fig. 1. Estimatedbody lengths of Accipitersand a few selectedcommon species of birds.For theAccipiters the verticalline indicates the mean (average)length; the tall rectangle includes one standard deviation .from the mean; the short rectangle, two standard deviations; the horizontal line, three standard deviations.

Volume33, Number 3 237 Goshawk

Cooper'sHawk

Sharp-shinned Hawk

I III

Fig. 2. Weightsof Accipitersand selectedspecies. Conventions as in Fig. 1. speciesare also taken from data col- the reader that this is true only when the respondents;one of the hawks was mis- lected at Cedar Grove, except for those tail is folded. We have found tail-shape identifiedby five of the sevenrespond- of the crow, which are taken from to be unreliable for identification. Five ents. Five of the six Cooper's photo- Roberts (1955). The average Common representativedrawings of Cooper'sand graphs were identified as Sharp- Grackle (Quiscalus quiscula) weighs Sharp-shinnedtails are shownin Figure shinnedsby one or more persons;one about the same as a d Sharp-shinned 3. We ask the reader to identify each Cooper's photograph was misidentified Hawk. Female Sharp-shinnedsaverage drawing before reading further; correct by everyone.The one Goshawk photo- slightlyheavier than kingfishersor flick- identificationswill be presentedbelow. graphwas called a Cooper'sHawk by six ers. FemaleCooper's Hawks weighabout The drawings are tracings made from persons and a Sharp-shinned by one. as much as crows but d Cooper's are photographsof birds captured and held The respondentswere told that there considerablylighter. Most Goshawksare in hand. The four photographswere might he a Goshawkamong the photo- at leasttwice as heavyas a crow. selected from 29 taken at our Cedar graphs;this resultedin two Cooper'sand The standard method of identifying Grove banding station in autumn 1978. one Sharp-shinnedbeing identified as birds is by using plumagecharacteris- These 29 photographswere taken of 26 Goshawks. If the above has not convinced tics;in identifyinghawks these are rarely birds: one each d and 9 Sharp-shinned you that the tail-shape of Accipiters is definitive, and often misleading. We Hawks, and one Cooper's Hawk were not a reliable character,compare your have tried to revieweverything that has photographedtwice: each with a folded identificationsof Fig. 3 with the follow- been written on the identification of North and then a slightlyspread tail. We asked ing: A, d Sharp-shinned;B. d Cooper's; American Accipiters. In the discussion sevengraduate studentsand professors, C and D, 9 Sharp-shinned;E, d that follows,we attempt to rectifythose all ornithologists, to identify the Cooper's.Figure 3 B is of a Cooper'stail publishedaccounts that are erroneousor photographs.All were permitted access with the tail feathers held somewhat misleading. to field guides. The averagescore was abnormally; but the appearance of a 75% correct. One of seven d Sharp- square tail would persist if the outer shinnedswas identified as a Cooper'sby feathers were coveredby the adjacent thatEVERAL Sharp-shinnedAUTHORS HAVE HawksINDICATED have oneperson. Nine of 11 9 Sharp-shinneds ones,which doeshappen normally. The notchedor squaretails. A few caution were called Cooper's by one or more squaredappearance is enhancedbecause

A B C Fig.3 (A-E).Tracings of photographs offive live, hand held Sharp-shinned and Cooper• hawks. The drawings have intentionally been drawti the same size.Identify each and compare your determinations with thecorrect ones given in thetext.

238 AmericanBirds, May 1979 the bird is in molt and several tail It has been suggestedthat Goshawks shinned,when movedforward in flight, feathers are not of full length, but tail have relatively longer wings than the may create the impressionof a leading molt is prolongedin adults and is often other two species.Our data suggestthat, edge which extendsfarther forward rel- found on birds during fall migration. If relative to body size, the wing of an ative to the head and body than in the the outer tail feathers are not of full immature d Goshawk averages 2% shorter-wingedCooper's Hawk. Relative length,this will createthe impressionof longer than an immature 9 Cooper's. to body size, the tail of Cooper's is a rounded tail; if central tail feathers are There is no difference between adults of 2%-3% shorter than in Sharp-shinneds, not of full length, this will create the the two species.Immature Accipiters of but this difference is insignificantwhen impressionof a notchedtail. all three specieshave wing lengthsthat compared to the difference in wing averageabout 1.5% shorterthan thoseof length. adults;so relativewing length appearsto C moreOOPER'S roundedHAWKS tailsDEFINITELY than do Sharp-HAVE be useless in identifying Goshawks. HUS,IN PROPORTIONTOBODY size, shinned Hawks (or Goshawks) but the However,if winglength is judged relative Sharp-shinnedshave longer wings differenceoften appears slight and is dif- to tail length rather than to body size, than the other two speciesand Goshawks ficult to discern.Female Sharp-shinneds Goshawks may appear longer winged have shorter tails than the other two havetails that appearmoderately rounded becauseof their relatively short tails. species.This leaves the Cooper's with even when folded and both males and Brett and Nagy(1973) and Haas (1975) relativelyshort wings and a long tail. females appear to have rounded tails have suggested that Cooper's Hawk Let us now turn our attention to differ- when the tail is spread in a fan, as in appearto havelarger headsthan Sharp- encesin plumagecolor and pattern. The soaring.Three of five personswe tested shinnedsor that the head of a Cooper's gray-barredbreast of the adult Goshawk identified a 9 Sharp-shinnedcorrectly projectsfarther in front of a line drawn clearly distinguishesit from the other when the tail was folded but called it a acrossthe leading edge of the wings. two species,but bewareof the occasional Cooper'swhen the tail wasslightly fanned. Examination of a number of museum pale-breasted9 Cooper's in poor light! Two of the sevenpersons tested noted specimensand living birds in the hand Many adult Goshawks have dark shaft that the codenumbers on the two photo- failed to showrelatively larger headsin streaks on the breast feathers; on some graphswere identicaland decidedthat Cooper'sthan in Sharp-shinneds.We individuals these are sufficiently prom- they were of the same bird. Tail-shape, think that it is possiblethat Cooper's inent so that the pattern of barring on at best, is an aid to identification but is may, on the average,appear to stretch the breast is considerably obscured. definitelynot a field mark. their necks a bit more than Sharp- Goshawks have a light-colored stripe Scheider (1975) and Haas (1975) have shinneds,but the way an individual bird overthe eyein all plumages,but in more noted that Goshawks have relatively holdsits headand wingsvaries consider- than a few immatures, particularly shortertails than the other two Accipiter ably from momentto moment.Further, males, this can be very inconspicuous, species.Analysis of our measurements we think the apparentprominence of the even with the bird in hand. Several hawk suggeststhat immature d Goshawks head in Cooper's, as compared to watchers have indicated that adult have tails that are 8% shorter (relative to Sharp-shinneds,is quite possiblya func- Cooper's have darker caps than adult body size) than immature 9 Cooper's tion of the relativelyshorter wing length Sharp-shinneds.Examination of a num- Hawks. The difference in adults is even in the former as comparedto the latter. ber of museum specimensreveals that greater: 13%. This difference, however, Relative to body size, the wings of the dark cap is a sex rather than a spe- must be used with care because adult Cooper's appear to be about 7%-8% cies related character. Adult d Sharp- Accipitershave 2%-4% shortertails than shorter than in Sharp-shinneds.We shinnedshave caps fully as dark as adult immatures. think that this long wing of the Sharp- d Cooper's.This characteris thususeful only if one has a sufficient idea of size to place the bird in the 9 Sharp-shinned/d Cooper's range, and get a rather good look at the cap.

HEBARRING ONTHE BREAST feather of an adult Cooper'stends to be of uniform width while in most Sharp- shinneds it is somewhat wider at the shaft. As a result the barring on a Cooper'sbreast appears, on the average, to be somewhatfiner and more regular than on that of a Sharp-shinnedHawk. At best, these characteristicsare merely other aids to identificationand require an excellent look at the bird. Peterson (1947) has indicated that immature Cooper's Hawks are usually moresharply and finely streaked below than immature Sharp-shinned Hawks. Examination of many museum spec- D E imens indicates that this would result in

Volume33, Number3 239 Although Accipiters are among the most difficult of the hawks to identify, many other speciesare also a problem. The systemof field marks and identifica- tion by eliminationsimply does not work with hawks, falcons, and eagles. We recall seeingthe followingspecies iden- tified as Sharp-shinnedHawks by other- wise highly reliable bird watchers: Cooper's, Goshawk. American Kestrel (Falco sparverius). Merlin (F. colum- barius). Broad-winged. As Cooper's: Sharp-shinned(very often!), Goshawk, Broad-winged, Red-shouldered (Buteo !ineatus), Red-tailed (B. jamaicensis), Peregrine(E. peregrinus).As Goshawks: Sharp-shinned, Cooper's, Red-tailed, Northern Harrier (Circus c.vaneus). We suspect that if we had kept a list of misidentifications we have witnessed Immature Goshawk. Photo/Harry IV. Darrow. over the years, it would include most of correct identification of about 80%-90% streaked. In immature Goshawks they the possiblecombinations of speciesof of the birds, if one had excellentjudg- are alwa.vs streakedor spotted. hawks and falcons. We urge extreme ment on the width and distinctness of The dark bars on the tail feather of an care in the identification of Accipiters, streaksand got a sufficientlygood look immature Goshawk are in the form of a and other diurnal raptors. One should at the bird. At the extremes, Cooper's V, while the dark bars on the tail feather not be too proud to list a bird as an have narrow streaks of nearly uniform of an immature Cooper's Hawk run "unidentified Accipiter" or even width and Sharp-shinnedshave streaks approximately straight across. This "unidentified hawk or falcon." which are distinctly tear-drop shaped. results in a zig-zag pattern of barring References Cited: Immature Sharp-shinnedsalso tend to across the Goshawk tail. contrasted with be somewhat buffier on the breast than a nearly even band in the tail of the BRETT, J. J. and A. C. NAGY. 1973. Feathers Cooper's,and the streakingstend to end Cooper's. This difference is easily dis- in the wind. Hawk Mountain Sanctuary somewhathigher on the belly in the tinguishedif one getsan excellentlook at Association.Kempton, PA. HAAS, F. 1975. Comments in Proc. North Cooper's than in Sharp-shinneds.but the barring in a spreadtail. The bands Amer. Hawk Migr. Conf. 1974.p. 123. both of these characters are highly appear irregular in the foldedtails of all PETERSON. R. T. 1947.A field guideto the variableand require an excellentlook at three species. birds. HoughtonMifflin Co.. Boston. the bird, and are therefore useless. To summarize the reliable plumage ROBERTS. T. S. 1955. Manual for the iden- tification of the birds of Minnesota and Several authors have indicated that characteristicsfor identifyingAccipiters, neighboringstates. Univ. MinnesotaPress, one should look for the "white, fluffy we have only the gray-barredbreast of Minneapolis. undertail coverts" on the Goshawk. If adult Goshawks,and the zig-zag barring SCHEIDER, F. 1975. Comments in Proc. this is intendedas a characterto identify of the tail of Goshawks. All other "field North Amer. Hawk Migr. Conf. 1974, p. immature Goshawks, we are mystified. marks" are qualitativecharacters which 122. WELTY, J. C. 1975. The life of birds. Second The extent, and "fluffiness" of undertail show overlap between speciesand pro- Ed. W. B. SaundersCo., Philadelphia. covertsis an apparent function of size, vide. at best, only a possibilityof correct but if "white" is supposedto indicate identification. --Departmen.t of Zoology and Curric- "immaculate" this statement is simply We end up then approximatelywhere ulum in Ecology, University of North false. The undertail coverts in immature we began:the bestway to identifyAccip- Carolina, Chapel Hi!l, N.C. 27514 Sharp-shinnedsand d Cooper's are iters is by sizeand size-relatedcharacter- (Mueller), Cedar Grove Ornithological almost alwaysimmaculate. In immature istics;a skill that can be acquiredonly Station, Rt. 1, Cedar Grove, Wis. 53013 •? Cooper'sthey are occasionallylightly with considerableexperience. (Bergerand A!!ez).

Onceagain American Birds will pub- the salonis published,if it is possibleto lish(if budgetaryconsiderations permit) frame the photographin the dimensions a competitivesalon of bird photographs of the American Birds cover. Each con- by its readers.Modest first, second,and testant will be limited to six entries in Ninth Salon of third prizes as well as honorablemen- each category;entries will be judged on tions will be awardedin two categories: the basis of technical excellence. artistic color and black-and-white. Color sub- merit, and difficulty. Pleaseidentify each missionsmay be in the form of trans- print or transparencywith name and Photography parenciesor prints.Deadline for accept- address,and furnish a stampedenvelope. ancewill be September1. First prize for Winners only will be asked later to sup- colorwill be honoredby being published ply biographical briefs and exposure as the front cover of the issue in which data on winningphotographs.

240 AmericanBirds, May 1979