Enemy Recognition by the Song Sparrow

Vol.•94• •8] J NmzAND T• PZLKWYK,Enemy Recognition bySong Sparrow 195

ENEMY RECOGNITION BY THE SONG SPARROW

BY MARGARET M. NICE AND JOOST TER PELKWYK

Plate 8

IN connectionwith the study of inborn and learned behavior in the SongSparrow (Melospiza melodia) an attemptwas made to experiment on fear reactionsin hand-raisedbirds, particularlyon the questionas to what it is that constitutesan 'enemy'to thesebirds. The SongSparrow is a favorablesubject for testsbecause the number of notes,either of 'alarm' or 'fear,' varieswith the degreeof excitement. In this we have an objectivecriterion of the degreeoœ fear. Hand-raisedGoldfinches (Spinus tristis), on the other hand, were either indifferentwhen tested,or flew wildly about the cage;it was easyto seewhat frightenedthem, but impossibleto make a com- parative study of the effectivenessof different stimuli. As stimuli in our tests we used live animals, mounted birds, and cardboard models. For the loan of the mounted birds we wish to thank Mr. E. R. Blake and Mr. Rudyerd Boulton of the Field Mu- seum of Natural History, Mr. Charles Rogers of Princeton Uni- versity,and Mr. Earl Wright of the ChicagoAcademy of Sciences.

OBSERVATIONS BY OTHERS

Severalexperimental approaches have recently been made to an analysisof the predator-preyrelation from the standpointof the predator (Tinbergen,1940); but often the assumptionseems to have been that the prey is passive.We know, however,that the prey has its specialmethods of defense,such as threatening,flight, hiding, or makinguse of protectivecoloration. Observationsand experiments on enemyrecognition are scattered. Bolles (1890) carrieda captive Barred Owl (Strix varia) into the woodsand recordedthe reactions of the birds towardit; the majorityof the birds "scolded"it, doing somore vigorously in summerthan in winter. Thomdike (1899)experimentedon instinctivereactions of young chicks,finding no in- stinctivefear of cats. "There developsin the first month a general fear of novel objectsin motion." A tame Carrion Crow (Corvus corone)and Jackdaw(Corvus monedula) were noted by Hertz (1926) to reactto everynew object with fear,but assoon as the objectproved 196 N•c•AnD T•R P•LawYa,Enemy Recognition bySong Sparrow [-AukL April

to be harmless,the fear changedinto curiosityand after that to ag- gressiveness.Various creatures were put into a cagewith a "hungry chickensnake" by Kellogg (1931); fear was shownby wild adult EnglishSparrows (Passer domesticus), but none by an adult Canary, nor by chicksless than a week old. The fear of snakesis often assumedto be inborn. Experimentsby Antonius (1939)show, however,that the well-knownsnake-fear in apesand monkeysis probably not inborn,but learnedfrom companions.No inborn fear of snakes wasfound by Rand (1937)in two hand-raisedBlue Jays (Cyanocitta cristata);a dogwas at first ignoredby the birds,but after one was38 daysold, it flew up out of the dog'sreach. Brfickner (1933) states that very young chickswere afraid only of loud noisesand of the lossof balance. Toward rabbits they behavedindifferently at first, but at the age of three weeksthey reactedwith distinct fear. There are a numberof experimentsusing birds of prey or models of the same. Lorenz (1935: 356) mentionsthe violent reactionof a hand-raised Passer domesticus to the first owl the bird ever saw; he statesthat the Magpie (Pica pica) instinctivelyrecognizes a preda- tory animal,but that the Jackdaw,that livesin flocks,does not. A mountedowl (Asio otus) and hawk (Accipiter)provoked fear in a tame Hooded Crow (Corvus cornix) and hand-raisedJackdaws (Strausz,1938). Plastermodels of the headsof thesebirds, painted or white, alsoprovoked fear when the headlooked in the direction of the Crow or Jackdaw. The birds were much lessdisturbed when the mountsand modelshad their backsturned. A mountedJack- daw arousedcuriosity or was ignored. Although the effectof the mountedowl and hawk was very great at first, after repeatedex- perimentsor if the objectswere left for sometime in the cage,the birds graduallybecame indifferent. Kr•itzig (1939)found in young of the Hazel Hen (Tetrastesbonasia rupestris) no signsof fear before the ageof ten days;after that theyshowed marked fear reactions to dogsand hawks. Goethe(1937), Tinbergen (1939),Lorenz (1939) and Kr•itzig (1940)describe experiments in which cardboardmodels of variousshapes were movedthrough the air alongwires. Goethe noted a specificreaction in young Blackcocks(Lyrurus tetrix) to modelsof birds of prey; there was a sexualdifferentiation in chicks as young as twenty days, the femalesseeking cover, the males as- suminga defensiveattitude. The experimentsof Lorenz and Tin- bergenshowed that youngGrey Geese(Anser anser) react from about the eighth week on; the form of the model was not as important with thesebirds and with ducklingsas with youngTurkeys (Meleagris THE AUK, VOL. 58 PI,ATE }q

Fro. I.--Attitu(les ()f' Song ."il)arrowin alarm. fcau and fright. From lef! to iigh! (Ul)l)el r(m-): I. miniarmed: '2. ale[t: 3. turning head to h)()k: I. 5. :illlB'in (l¾illg•,alld tail fill)peal): strong alnrm; 7. fear: 8. fright (panting).

Fit;. 2.--Cardl)oard models of oxvls. I"lom lett I{) right: I. model at half scale; 2, 'head' model: 3. 'l)csl' model; 4, 'head' model xxilhmtt c,xcs;5, 'outline' of 'head' model; 6. model with shading on the sides. Vol.194158'] .J NICEAND TER PELKWYK, Enemy Recognition bySong Sparrow 197

gallopavo)and Kr/itzig'sRock Ptarmigans(Lagopus mutus). With both thesespecies a modelwith a shortneck and long tail (Acci- piter) provokedstrong fear reactions,but the samemodel was largely ignoredwhen pulled overheadwith the long end forward (goose). Young hand-raisedptarmigans were frightenedby a live dog, but paid no attention to modelsof mammals. Kelso (1940) found that youngScreech Owls (Otus asio)showed fear at the sightof the skin of a Crow (Corvusbrachyrhynchos) when it wasmoved. In a letter Mrs. Amelia Laskeydescribes some observations made on a hand-raisedMockingbird (Mimus polyglottos):"Jan. 11, he fights everythingthat comeshis way including the dog, his shadow,his reflection,and other birds that I hold near him exceptinga dead ScreechOwl. To that he givesa pronouncedfear reaction,flying in frightened,fluttering manner as far back in his cageas possible when he seesthe owl. He attackeda SparrowHawk throughthe wire." On June 11, she wrote: "When I had the Saw-whetOwl (Cryptoglauxacadica), he showedfear of it. One eveningthe owl was free in the room and flew to a perch abovethe Mockingbird's cage. The Mockingbirdcraned his neck repeatedlyto watch the owl who wasfour feet abovehis cage,giving sharp staccato chi-chick callsrepeatedly. The owl bent its head a numberof timesto look at the Mockingbird,but showedno specialinterest even when the room lights were dimmed. After the owl was taken out of the room all lightswere switched off. In aboutten minutesthe Mockingbird beganto give the staccatocalls again. I went quicklyto the room, turned on bright lights and saw the Mockingbirdcrane his neck and turn his headsideways looking toward the spotwhere the owl had perched." A curiousphenomenon has been mentionedby the Heinroths (1924-32),namely, that somebirds are afraid of colors,some of blue, others of red. Among the former they cite one of a brood of five Red-backedShrikes (Lanius collurio), three of a broodof five Gray Wagtails (Motacillafiava), a Jackdaw,a Yellowhammer(Emberiza citrinella) and a Cockatoo. Tree Pipits (Anthus trivialis) were afraid of red. A pair of Goldfinches,taken from the nest at the age of two days by Wm. E. Schantzand later brought to Chicago, were afraid of a dark-bluebox top, flyingwildly in the cagewhen- ever it was shown. They fearedsome other blue objects,but not miscellaneousobjects of other colors. Another pair raised by Mr. Schantz,showed fear only to black paper. In our experimentswith SongSparrows we did not observea fear for any color. 198 NInEa•D T•et• Pr_J•KW'VK, Enemy Recognition bySong Sparrow I-Auk1 April

Bm•Av•oR oF XHE SONG S•'•ROW Three chiefstages of fear maybe distinguishedin the SongSparrow: Stage Note Postures tchunk Crest raised; tail raised and flipped; wingsflipped; restlesschange of location. tik Featherscompressed; neck elongated; body crouched. 'Fright' tik-tik-tik Flies and hides; flutters in attempt to escape;pants with open bill.

These categoriesshow 'alarm,' 'fear' and 'fright' in their extreme form; in milder casessome of the characteristicsare not shown. In the first stageall characteristicsexcept the raisedcrest may be dis- playedseparately in situationsthat are obviouslynot concernedwith alarm. The note tchunk is the 'ordinary' note of the male Song Sparrowon his territory,and is seldomheard from the femaleex- ceptin the nestingseason; it is typicallygiven by her directlyafter shehas left the nestduring incubationand more vigorouslyin con- nection with the approachof an enemy to the nest. It often seems to have an element of 'protest' in it. Sometimesit seemsto be givenby the male as a matter of 'self-assertion.'(With the hand- raisedmales it is often given while bathing; here it might be inter- preted as vocal self-assertionwhile temporarilyhampered by wet feathers.) It appearsto be analogousto the 'rain-call'of the Chaf- finch (Fringilla coelebs),as describedby Sick (1939), which often expressesdiscomfort or anxiety,and alsoappears to be a specialex- pressionof territoryadvertising. Wing- and tail-flippingmay accompany'pleasurable' excitement, as when the pair hunt for a nestingsite. Raising the crestnearly alwaysappears to expresssome alarm or apprehension.It hasbeen characteristicof the underling male with the hand-raisedbirds. The rhythm with which the notesare givenand the speciesof note vary with the degreeof excitement.To illustratethis, three ex- ampleswill be givenwith wild SongSparrows in Massachusettsin 1940. June 12: the observerin searchingfor a nestwhich contained newlyhatched young came within a few metersof it; the female gave48 tchunksand five tiks in one minute. The observerleft, walked 50 metersand hid; the bird then gave 32 tchunksin a minute. Two days later when the nest was visited, the female 'gave56 tchunksin a minute;as soonas the observerwas out of sightshe gave 35 in a minute. ,' Vol.•8]194! .I NitreAs• T•a•Pv_sacw¾•, Enemy Recognition by Song Sparrow 199

July 11: one parent was 'scolding'a cat, 8 metersfrom the nest containing6-7-day young; at first the cat waslooking at the Song Sparrow, later it turned its back and chewed a weed. In three consecutiveminutes the bird gave 51 notesa minute: in the first minute 1 tchunk and 50 tiks, in the next 15 tchunks and 36 tiks, in the third, 47 tchunks and four tiks. June 30: a mountedBarred Owl wasfixed near a tanglein which were a Song Sparrow and young that had recently left the nest. The adult gave 64 tiks in the first minute, 63 in the next and 65 in the third. The first bird showed alarm: strong alarm with 56 tchunks a minute and moderate alarm with 32-35 a minute. The second bird showedfear first, later alarm: fear with approximately51 tiks a minute, and alarm with approximately51 tchunks. The third bird showed fear with 64 tiks a minute.

REACTIONS OF SONG SPARROWS IN NATURE TO ENEMIES in a letter Richard Pough informs us that the chief enemiesof SongSparrows among the hawksand owlsare, in orderof importance: Cooper'sand Sharp-shinnedHawks (Accipiter cooperi and velox), Long-earedand ScreechOwls (Asio wilsonJanusand Otus asio), Pigeon and SparrowHawks (Falco columbariusand spa;verius), Marsh Hawk (Circus hudsonius)and Red-tailed Hawk (Buteo borealis). Sherman(1911) found that Juncos(Junco hyemalis) and SongSparrows were the most frequentvictims of a nestingpair of Screech Owls which she studied in Iowa. SongSparrows were studiedfor nine yearsin the field on Inter- pont, Columbus,Ohio (Nice, 1937); the enemiesmost frequently recognizedby the birds were the cat and the Cowbird (Molothrus ater). Other enemiesthat might have taken toll of the SongSpar- rowswere: snakes, dogs, rats, opossums, weasels, skunks, red squirrels, chipmunks,Pheasant (Phasianuscolchicus torquatus), Bronzed Grackle (Quiscalusquiscula aeneus), Blue Jay, Accipiters,Sparrow Hawk, Screech Owls. Reactions of the birds to a number of enemies were observed. Snakes.--Alarge garter snakelying 60 cm. below 1M's nest with newlyhatched young was treated with curiosityby 1M; he hopped aroundit and gaveit a peck. The following.year 1M's mate on five occasionswas seen to attacksmall snakes in the vicinityof her nest containingyoung; sometimes she was silent, but twice she gave sputtering notes. 1M was present in one instance,but remained indifferent. 200 NIcI•ANO TIeR I'EI2•WYK, Enemy Recognition bySong Sparrow I'Auk/ April

Rabbits.--Cottontailrabbits were consistentlyignored, except that oncea femalegave the threat note to a youngrabbit sittingnear her nest containingeggs. Cats.--Catswere consistentlygreeted with tchunksgiven at a rapid rate (strongalarm), the tchunksbeing taken up by all the SongSpar- rows in the neighborhood. Once parentswith young ten days out of the nest gave tik-tik-tik at a cat. Dogs.--Dogswere ignored unless they approachednear the nest, when tchunk was given. On severaloccasions dogs broke up nests, killing the female and (or) young. People.-Usually people are greetedwith tchunk when they come near the nest. When the young are about ready to leave the nest and shortly afterward, the parents,especially the female, are apt to usetik aswell astchunk and alsoa varietyof •iscellaneousnotes, that are only usedin connectionwith disturbanceover young. Oc- casionallytchunk is usedin the winter whena SongSparrow has been startledby the suddenappearance of a person.•Tik-tik-tik has been recordedfrom SongSparrows in relation to peoplewhen the birds are in a heightenedstate of excitementover mate or young. It was given by a male a few daysafter being joined by a mate, and by 4M when both his first and his secondmate in 1935 were trapped and the bander came to get them. It was given by 5M and 4M with young two daysout of the nest, and by the latter when one of his young, a week out of the nest, came near the observer. It was also given by a femalewhen a personapproached her well-grownCow- bird foster-child. Cowbird.--Duringthe nestingseason Song Sparrowsreact to the presenceof the femaleCowbird on their territorieswith emphatic tchunk-ing. If the Cowbirdcomes near the nest,the pair attackher. One pair wereseen doing this in the presenceof one of their young that wassix weeksold. Recordsof alarm over the presenceof Cow- birds range from March 30 to July 14. A male that was known to be a juvenilebird by the characterof his singingreacted strongly from April 18 on to the visitsof courtingCowbirds to the chief tree in his territory. Ovenbirds(Seiurus aurocapillus), although heavily parasitizedby the Cowbird,do not recognizethis bird as an enemy (Harm, 1937). BronzedGrackle.-T•wice individuals of this specieswere observed near SongSparrow nests, the parentsof which were reactingwith strong alarm. Robin.--Onone occasion Song Sparrows were tchunk.ing at a Robin (Turdus migratorius)near their nest. Smaller birds in such'situations were driven off. , Vol.•94• 58] a NICEta•v Tz'• PExacwY•c,Enemy Recognition bySong Sparrow 201

SparrowHawk.--Once (March 15) the SongsSparrowsgave tik-tik- tik when a pair of Falco sparveriusappeared; on anotheroccasion (November2) they gavetchunks; twice (Februaryand October)they fell silent,and finally, when one of thesefalcons flew at two fighting males,they hid. In the vicinity of the falcon'snest the Song Spar- rows were indifferent and this was also true in general throughout the nestingseason. Three times a falcon was seen trying to catch a small bird without success,but twice individuals were seeneating small birds. Marsh Hawk.--All SongSparrows stopped singing and many said tik-tik-tik when a large female flew over slowly on March 1. Sharp-shinnedHawk.--The Song Sparrowsgave tik-tik-tik, hid, and remained 'frozen' for some minutes. Red-shoulderedHawk.--On August 17, 1940, in Pelham, Massa- chusetts,a Buteo lineatuswas seen to fly over a swamp;a few minutes later the observerreached this locality and heard loud tchunksgiven rapidly by a male SongSparrow seated on top of an alder. In four differentminutes from 12.25 to 12.36he gave the followingnumber of tchunks a minute: 108, 92, 76, 64. His mate in the meantime gave 27 tchunksand five tiks in one minute, 24 tchunksin another. The observerleft the swampand at 12.40 the male gave 18 tchunks in one minute. The pair had youngrecently out of the nest; their excitementwas undoubtedlydue partly to the very recent sight of the hawk and partly to the presenc•of the observer. The male was unusualin giving tchunksat sucha very rapid rate insteadof using tiks; the next highestrecord we have for number of tchunksper minute is 70 from Y. Owls.--There was no opportunity to observethe reaction of wild SongSparrows to thesebirds, except the testalready mentioned with the mounted Barred Owl. Reaction of small birds to owls, cats, and some other enemies has a biologicalfunction; the 'alarm' notesare understoodby birds of many other species;a number gather and, following the enemywith loud cries,make it impossiblefor it to come unheraldedupon its prey. EXPERIMENTS WITH ADULT HAND-RAISED BIRDS

The subjects.--Mostof the experimentswcrc done with the male Y, but someresults wcrc obtained on the male D and the female J before their deathsby accidentin March 1940. Y was taken from the nest in Pelham, Massachusetts,on August 1, 1938, at the age of five days. D wastaken from the nestin Augusta,Michigan, on June 202 N•cEANO Tr• PI•LKWYK,Enemy Recognition bySong Sparrow LI'Auk April

20, 1939,at the ageof six days. J wastaken from the nestwhen about eight daysold in Columbus,Ohio, August 12, 1939,by Mr. W. E. Schantzand broughtto Chicagoin November. The maleshad the liberty of the study until December 1939, when it becamenecessary becauseof territorial fightingto keep one or the other caged. J was kept in a large cage60 by 90 by 120 cm. in size,later usedfor Y in the experimentscarried out in May 1940. D and Y were always experimentedon separately,but this was not true with J. She was presentwhen Y wastested with the mountedBarred Owl on February

CHARTl.--Diagram showing the numberof tchunksper minute givenby Y to threemounted birds: Barred Owl, BurrowingOwl, falcon (Micrastur),each shown for 15 minutesat aboutfour metersfrom the cage. This illustratesthe decreasein the reactionduring an experiment.

1 and 25, when he was testedwith the teddy-bearon February25, and whenD wasshown the teddy-bearFebruary 9. Experiments.--Fiveseries of testswere carried out: 1. Live animalsbrought within 1.5 metersfrom the cageor even releasedinto the cage:snakes, rat, rabbits,dogs, cats. 2. Movingmodels: mounted Barred Owl with movingwings, flying modelsoutside the window,models brought suddenly to the cage. 3. Seriesof objects,mostly mounted birds, each shown for 15 minuteson a victrola about four metersfrom the cage:Barred Owl, hornbill, Ruffed Grouse,Burrowing Owl, falcon,kite; also a teddy- bear and a largepitcher (Chart 1). 4. Seriesof objects,mostly cardboard, shown on the victrolawith white cardboardas a background;each shown for three minuteswith Vol.x94t•8'[ • N•c•Am) T•R PELKWVK, Enemy Recognition bySong Sparrow 203 five-minuteintervals between tests. The models(Plate 8, fig. 2) were largelyof owls,but includeda duck,quail, hawks, and a shorebird. In eachexperiment one object was used as a standard,and reshown at intervalsto test the level of the bird's reaction. The sequenceof the modelswas such that strongand weakstimuli alternated, so that everytest showed as much as possible the reactionto that particular model (Charts2 and 3).

C•tnRT2.--Diagram showing the numberof tchunksduring the three-minutetests (A) andfive-minute intervals (B) on May19. The samemodel ('head' model) was shownafter every three tests, giving a standardof reactionabilityduring the experiment. 1, 5, 9, 13,'head' model (Plate 8, fig.2, no.2); 2, duck;3, kite;4, 8, 12,'outline'of 'head'model (Plate 8, fig.2, no.5); 6, shorebird;7, DuckHawk; 10, model with shadingon the sides(Plate 8, fig.2, no. 6); 11,quail.

5. MountedBarred Owl on thepiano in thefront room. Alsothe same situation without the owl. The seriesof testsfalls into two groups:the two firstare experimentswith movingenemies, the threelast with stationaryenemies. In interpretingthe reactions of oursubjects we use the criteria for 'fear'and 'fright'mentioned previously. As to 'alarm,'in general we found the rate of the notesfell into three fairly distinctclasses: 10-15 a minute (weakalarm), 25-30 a minute (moderatealarm), andabout 50 a minute(strong alarm). In threeminutes the number of tchunksin weak alarm rangedfrom about 40-65, in moderate 20z• NIcœAND TER PEI•WYK, Enemy Recognition bySong Sparrow kF AukApril

C•,•RT3.--Diagram showing number of tchunksper minuteduring one seriesof tests (May 17). The time and duration of the testsare representedby lines at the top of the chart. a, representsthe 'best'model (Plate8, fig. 2, no. 8); b, the 'outline'of the 'head'model (Plate8, fig. 2, no. 5); c, the 'head'model (Plate8, fig. 2, no. 2): d, the 'outline'of the owl modelat half scale;e, the owl modelat half scale (Plate 8, fig. 2, no. 1); [, the 'head' model upsidedown.

alarmabout 80-120, and in strongalarm about 130-200. The three- minute records are not as clear cut as the one-minute ones, since Y would seldomstart tchunk-ingthe first momentthe experimentbe- gan, and sometimeshe did not start for a full half-minute or more; in these cases the derision as to whether the reaction should be classedas moderateor strongalarm had to be basedon the number of tchunksgiven in the secondand third minutes. Sincethe majority of the testslasted three minutes,we use this period as our standardin the followingreport and unlessother- wise statedthe number of tchunksmentioned was that given in three minutes. MovingEnemies In generalthe reactionsare strongerto moving than to stationary objects. The birds had never beforehad experiencewith the first three animals shown in the tests, so we conclude that the reactions are partly inborn. Snake.-Two garter snakes (Thamnophis sirtalis) 30 and 70 cm. in length were shown within one-meter distance. Both moved rapidly and the smallerone was allowed in the cage. Y reactedby craninghis neck,but showedno signsof alarm. Rat.-Two activewhite rats,one half grown,the otherfully grown, were broughtto Y's cageand the smallerone allowedin the cage. Y reactedby elongatinghimself, the characteristicattitude of curiosity, but showed no alarm. Vol.1941sS' .•I NIcv.A•I• Tga Pv.•.•w¾•, Enemy Recognition bySong Sparrow 205

Rabbit.--A large black-and-whiterabbit was placednext Y's cage, where it hoppedabout. Y flew back and forth high in the cage, pantingwith open bill (fright), but gaveno note. The rabbit was removedand shownagain in the sameplace after eight minutes;at thistime it wasquiet. Y gave107 tchunks in threeminutes (moderate alarm). Dog.--On April 1 a medium-sized,active, white dog was brought within a meter of the cage;Y flew about as high as he could (fright), but gave no note, until two minutesafter the dog wasremoved, when he gave25 tchunksin one minute. On September18, a youngchow was brought to the cage;Y flew wildly back and forth; the dog was removedand Y gavea few tiks. In a day or two Y becameindifferent to the puppy. ½at.--No reactionwas ever given to catsseen out the window. On September17, 1939,three small kittenswere brought in and placed near the cage;Y reactedwith moderatealarm. On April 15, 1940,a medium-sizedblack cat washeld one and a half metersfrom the cage, where it moved somewhat,purred and coughed. Y reactedby raising his crestand flipping his tail and gave the following number of tchunksduring the five-minutetest: 10, 20, 12, 0, 0 (weak alarm). During the summer Y was cared for by Mrs. Dorothea Ewers who kept him on an outdoor screenedporch; on severaloccasions Mrs. Ewersheard him giving tik tik, whereuponshe went out and chased a cat from the railing. On September23, a smallgray-striped kitten was placednext his cage. Y was disturbed,flying back and forth high in the cage,but gave only six tchunks. To a small white kitten he behaved in a similar manner,but gave 72 tchunksand three tiks in three minutes. Birds outside the window.--Sometimesthe Song Sparrowsshow fright reactions,giving tikotikotikand hiding, apparentlystimulated by the sightof somebird in the sky. On May 27, a Starling(Sturnus vulgaris) flew rapidly by at 10 to 15 metersdistance; Y gave the tik-tik-tik note, flew to the side of the cageand hid (fright). On September27 a Blue Jay appearedoutside, screaming; Y gavea series of tiks for about a minute. Mounted Barred Owl with movablewings.-On May 16 when the modelwas stationary,Y showedstrong alarm (133 tchunks). After a five-minuteinterval the owl was again shown,but this time the wingswere movedup and down. Y showedfear (94 tiks and 20 tchunks)and tried to escape. On May 27 with the modelstationary Y gave 113 tchunksand three tiks; then the wingswere movedand Y showedfear with 70 tiks in one minute, in the meantimeflying back and forth in the cage. 206 NxcEA•D TER PE•CW¾•C, Enemy Recognition bySong Sparrow ['Auk[.April Flyingmodels.--Flying models were shown outside the window, onein the shapeof an Accipiter,the othera rectangle.At the first sightY showedmomentary fear. As long as the modelsmoved slowly,there was no distinctreaction. If the modelswere rapidly moved,Y showedfear at both. Apparentlysudden movement and not form frightenedhim. Modelsbrought suddenly to the cage.--Severalobjects were shown suddenlyand with movementfor about four seconds.They were presentedin the followingorder: mountedBurrowing Owl, rec- tangularpiece of cardboard,small cardboard model of owl (Plate8, fig.2, no. 1),unpainted side of thesame model, painted side, piece of cardboard. Y reactedwith fright to the mounted owl and the paintedside of the model,but wasattentive and, during the lasttest, slightlyalarmed by the pieceof cardboardand the unpaintedmodel. StationaryEnemies In noneof the experimentswith the stationarymodels did the birdsshow any greaterreaction than strongalarm. All experiments were characterizedby a gradualdecrease of the reaction. This is well illustrated in charts 1 and 2. MountedBarred Owl.--On February1, the mountedBarred Owl wasshown in the study;Y wasstrongly alarmed, while the femaleJ elongatedherself. On February25, it wasdisplayed on the victrola; Y wasstrongly alarmed; with tail flipping and crestraised, he gave about nine tiks and 60 tchunksin a minute. J raisedher crestand flirted her tail (shenever gave the notestchunk or tik). After five minutesthe owl wasremoved, but Y gave about 36 tchunksa minute for the next two minutes. From April 2 to 12, testswith mounted birds were given every other day; on the latter date the owl was shownand evokedonly moderatealarm with 114 tchunksin the first three minutes; however, there was no cessationof tchunks during the 15-minuteexperiment, 405 being given in this period; of the eight objectsshown it was only with the owl and the hornbill that tchunkswere given throughout the 15 minutes. On May 15, the owl was used as the standard in the series of three-minute tests; the number of tchunksgiven showeda marked decrease,namely, 138, 102, 73. Interestinglyenough, this mount invariably evoked strong alarm in both Y and D when it stoodon the piano in the front room. This was true on five occasionsfrom February 6-15 when D saw it in this situation and six times between March 23 and May 24, when Y had this experience. For example,on April 12, Y went into the Vol.•94• •8] .I NIcE.•v Tr• P•r•.KWX'K,Enemy Recognition bySong Sparrow 207

middle room, reachedthe point from which he caughtsight of the owl, and with crest up and tail flirted, gave 202 tchunksin 225 seconds,and flew back into the study. On four of Y's six visits an exact record of tchunhswas kept; the countswere high: 166, 158, 214, 161. Apparently the birds were conditionedafter theseexperiments to the situation,for they showedalarm merelyat the sightof the piano, when the owl was absent. D showedstrong alarm on the afternoon of February 15, when he found the owl absentafter six visits during the previousnine dayson all of which the owl had been present. On April 4, Y reacted with moderate alarm to the absenceof the owl, after havingseen it March 25 and April 1. His reactionto its reappearanceon April 12 has already been described. On his next visit, May 8, the owl had beenremoved; he enteredthe room keeping his eyeson the spot where the owl had been displayed,and showedstrong alarm, giving 180 tchunhsin the first three minutes, after which he fell silent, graduallygave up signsof alarm, and nine minutes after entering the room began to sing softly within three metersof the piano. The owl was presentat his next three visits (May 20-24), but absent May $1; on this occasionY gave the followingnumber of tchunksper minute: 49, $8, 22, 15, 5, 2, after which he returnedto the study. After more than four monthsY wasstill conditionedto this spot. During most of the summerhe had lived in another house. He had not seen the owl on the piano since May 24. On October 1, he enteredthe front room, at onceshowed strong alarm with crestand tail raised,keeping his eyeson the end of the piano. He gave148 tchunksin three minutes,left the front room and gave nine more beforehe returnedto the study. In the fall we useda differentmounted specimen and found the reactionof Y to this specimenmuch weaker than it had beento the first specimenor to the cardboardmodels of owls. On October 11, at the endof a shortseries of testshe showedonly weak alarm (44 tchunks)to themount, although the 'best' cardboard model (Plate 8, fig.2, no. $) wasas effective as in the spring(115 tchunks). In our eyesthis secondmount seemsa poor specimen.On October16, we placedthis owl on the piano;Y, on enteringthe room,raised his crestand flirted his tail and gave42 tchunksin the first minute and 45 in thesecond, after which he ttewback to thestudy where he gave 13 more tchunks. Other mounted birds.--Tests were made with a number of mounted birdsfor the sakeof comparisonwith the reactionsto the mounted 208 NIcEA•D T• ?I•LI•WYI•,Enemy Recognition bySong Sparrow •AukL April

Barred Owl. In generalthese birds causedless-marked alarm. A mountedhornbill, about 40 cm. long, grayishin color, evokedthe strongestreaction of any of thesespecimens, causing strong alarm in Y at the firsttest (154tchunks), and moderatealarm (117 tchunks) at the second.A kite (½hondrohierax)caused moderate alarm (123 and 118 tchunks). A falcon (Micrastur)and a BurrowingOwl (Speotyto)evoked moderate alarm (93-109 tchunks),although in one test (May 22) Y showedno reactionto this owl, which,it should be noted, was a decidedlypoor mount. A Ruffed Grouse (Bonasa umbellus)gave inconsistentresults: strong alarm at the first test (April 1) andweak (47 tchunks)five days later; on May 22, it evoked 70 tchunks in the first test and none in a second test ten minutes later. Observationsin Nature show that Song Sparrowson their terri- toriesreact with strongalarm to the visitsof female Cowbirds. On September26, a mountedfemale Cowbird was placed30 cm. from Y's cage;Y gave 83 tchunks(moderate alarm). On October6, we comparedthe reaction to the mounted Cowbird with that to a mounted Starling (Sturnusvulgaris); Y showedweak alarm to the former (37 tchunks)and moderatealarm to the latter (90 and 116 tchunks). It seemsprobable that the strongalarm shownto Cow- birdsin Nature is basedon experience. Cardboard models.-Since the reaction to the Barred Owl was so much more pronouncedthan to the other birds, we askedour- selveswhat the differencemight be to the Song Sparrowbetween the owl and the other birds. In an attempt to answerthis question we made a seriesof models,cut out of cardboardand painted on one side with India ink and crayon. These modelswere shown on the victrola in three-minute tests. The 'best'model (Plate8, fig. 2, no. 3) wasas far aspossible a copy of the mounted owl. Y reacted to this model as much as to the mounted owl in the same situation,showing for the most part moderatealarm (94, 115; 148, 128, 102, 124; 103, 97, 98; 121; 116 tchunkson five different days). The same model was shownwith the eyescovered; Y gave 93 tchunks. A pieceof plain cardboard of the sameshape was shown; Y gave 16 tchunksin the first test and one in the second. To find out whether the size of the owl was important, we made a reproductionof the 'best' model at half scale (Plate 8, fig. 2, no. 1). Y showedmoderate alarm with 92 tchunkson May 17 and weakalarm with 34 tchunksfive dayslater. To a piece of cardboardwith the sameoutline Y gave no reaction. The model we usedmost often asa standardwas the 'head'model (Plate8, fig. 2, no. 2), a reproductionof the head of the 'best'model with a slender Vol.104158'[ .• N•c•A•;D T• PgL•w¾•,Enemy Recognition bySong Sparrow 209

unpainted body. This model proved to be almost as effectiveas the 'best'model, evokingmostly moderate alarm (119; 133, 107, 94, 114; 97, 92, 94; 152, 125, 114, 68; 83, 72; 53; 72; 91, 10 tchunks). These series,obtained on eight days betweenMay 15 and 27, il- lustrate the decreasein the intensity of the reaction during this period. Two testsmade in the fall showedmoderate alarm (93 and 97 tchunks). A piece of plain cardboardof the sameshape as the 'head'model (the 'outline') (Plate 8, fig.2, no. 5) provokedno reaction in nine testsand very weak alarm (14 tchunks)in one. So there is a remarkabledifference between the resultswith the painted and unpainted models. We attemptedto analyzethis differenceand in this way to find the essentialfeature in the owl. The 'head' model without eyes (Plate 8, fig. 2, no. 4) evokedmoderate alarm (104 tchunks). The 'head'model placed upside down brought 76 tchunks. Reactionswere similar to a modelof which only the left half was painted (77, 82 tchunks). A model with only someshading on the head evoked53, 27 and 0 tchunks,and when eyeswere addedno reactionwas given. A model with only someshading on the breastbrought 13, 66 and 0 tchunks. A model with a dark border around the outline evoked no reaction,but a similarmodel with someshading along the outline, so that it seemedto be rounded (Plate 8, fig. 2, no. 6) brought94 tchunks. A pieceof cardboardof the samesize as the 'head' model wasgradually changed and shownat intervals. Eyes,bill and disks aroundthe eyesbrought no reaction,but as soonas someshading wasadded that gavecontour to the model,Y showedmoderate alarm, uttering 85 tchunks. Similar models were made of various birds: duck,kite, duckhawk, quail and shorebird;all of thesecaused some alarm (67-99 tchunks).Tests on October11 gave the following figures.'shorebird, 64 tchunks;duck hawk, 118, 99. Other objects.--Inorder to interpretthe resultsobtained with the mounts and models, it is of interest to know how the birds react to new objectsin their environment. Ordinarily the hand-raised SongSparrows have been indifferent to new objectsbrought into theroom. Someobjects, however, evidently have a specialmeaning, ashave some of thecardboard models. On February9, a teddy-bear about30 cm.high wasplaced a meterand half from the cage;D showedstrong alarm (33 tchunksin 30 seconds),while J raisedher crest. In similartests Y showedmoderate alarm on April 4 (108 tchunks)and weak alarm on April 15 (58 tchunks),but on October$ he reactedwith strongalarm (127tchunks) and the next daywith moderatealarm (84 tchunks).To a 20-cm.-highpewter pitcher Y ANnTER PELKWYK, Enemy Recognition bySong Sparrow [Auk!. April

reactedwith moderatealarm for two minutes,giving 67'tchunks, after that becomingindifferent. Notes of owls.--Birdsin Nature sometimesrespond to notesof hawks and owls and even to imitations of these. On October 11, Mr. J. Murray Spiersgave excellent imitations of the hootsof Great Horned, Barred and Barn Owls and also of the cry of the Red- shoulderedHawk, without eliciting any responsefrom Y. Experimentswith Goldfinches.--Afew experimentswere made with two hand-raisedGoldfinches when they were one year old. A series of objectswas shownbut no reaction was given to the teddy-bear, nor to any of the cardboardmodels, but the mounted Barred Owl evokedfright, both birds flying rapidly about. After a five-minute interval the 'best' model was shownagain; now the birds flew back and forth. Somewhat similar reactions, but weaker, were shown to the other models,the male evincingmuch more concernthan the female. With this speciesit is difficultto measurethe degreeof excitement.

EXPERIMENTS WITH YOUNG SONG SPARROWS During threesummers Song Sparrows have been raised by hand, beingtaken from the nestat the ageof fiveto sevendays. Situations to whichthese birds instinctively responded with escapereactions were: threatenedcapture, movement of the immediateenvironment, approachof very large movingobjects, evidence of fright in companions,sight of birds flying overhead,the specificfear note, and soundsresembling it. Duringtheir fourth week one brood (Y and hisbrothers) did not recognizea cat or a dog as an enemy;the samewas true with the maleA when75 to 80 daysold. A hornedtoad (Phrynosoma)and groundsquirrel (Citellustridecemlineatus) to which the broodwere introducedat the ageof one monthevoked curiosity. During 1940,a mountedBarred Owl wasshown to four young birds:to threeof themat intervalsfrom 7 to 21 days,to A at 7, 17, 21, 28, 32, 37, 50 and 62 days. Three of the cardboardmodels of owlswere also shown during the last threeexperiments. All tests gave negativeresults. When A was three months old he was shown a seriesof modelsand the mountedCowbird, Starling and Barred Owl. To none of thesedid he react exceptin two instances:he gave 61 tchunks in three minutes to the outline of the owl with the darkshading (Plate 8, fig.2, no.6), but in a secondtest gave no reaction;to theCowbird he gave 52 tchunksat thefirst showing, but no re- actionin thesecond. A hasbeen shown the owl once a monthup to Vol.194!S$- -•I NiceASP TER PELKW'/K, Enemy Recognition bySong Sparrow 211

the age of sevenmonths and is still indifferent. UnfortunatelyA hasalways been a sub-normalbird; this is clearlyshown by the fact, that, althoughhis wing measurementshows him to be a male,he has neversung, while the four other hand-raisedbirds sang a greatdeal in their first fall and winter. The experimentswith the adult Song Sparrowscannot be ex- plained without postulatingan inborn pattern, but possiblythis pattern appearslate.

DISCUSSION Apparentlythe SongSparrow recognizes enemies by both inborn and learned patterns. In Nature the behavior of the adults is of great importancefor the formingof conditionedpatterns in the young. Simultaneouspresentation of a nativelyinadequate stimulus (cat, Cowbird)and the unconditionedstimulus (alarm of the adult)may result in conditioningto theoriginally inadequate stimulus. Evidentlyinborn is enemyrecognition of large,new, moving objects,and specificallyof an owl. Evidentlylearned, either from personalexperience or from reac- tionsof others,is enemyrecognition of cat and Cowbird. On the otherhand, the birdsmay learn in one or both wayswhat is not an enemy. The experimentsshow decrease of inborn reactionwhen nothinghappens after the enemy has been shown. A similarprocess may be involvedin the absenceof alarm shownfor rabbitsand often for SparrowHawks (Falcosparverius) in Nature. Conditioningwas certainly an importantfactor in the experiments. Whydid thesituation in thefront room bring such a marked reaction?One of thefactors was undoubtedly the familiarityof the environmentin the studyand the unfamiliarityof that in the front room. Mostof Y'sexperiences in the studyhad nothing to do with enemies,but after March 23 theowl was the most important element of the front roomfor Y. Lorenz (1935:205) describeshow his Ravens(Corvus corax) were conditioned against places where they had beenfrightened. The experienceof Mrs.Laskey shows how quicklya Mockingbirdwas conditioned to the placewhere an owl had perched. Althoughthe victrola evidently had a neutraltone when displaying indifferentmodels, there was, however, probably some influence from the owl that was reflected in the reaction to other mounts and to someof the models. It is not clear if the characteristicsof the owl pattern-head and shading-are characteristicsof the inborn owl patternor merelyelements of previousexperience. 212 NICEANO 'I'ZR PELKWYK, Enemy Recognition bySong Sparrow [AukI. April

SUMM•,•Y 1. A surveyis given of previousobservations on the relation 'enemy-prey'from the viewpointof the prey. Fear reactionsto owls have been reportedin hand-raisedindividuals of Passerdomesticus, Corvusmonedula, Mimus polyglottos,and to hawksin Corvusmo- nedula, Tetrastes,Lagopus, geese and ducks. 2. In the Song Sparrow it is possibleto differentiatestages of 'alarm,' 'fear,' and 'fright.' 3. In Nature, the most important enemiesof the Song Sparrow are probablycats, hawks and owls. 4. Catsarouse strong alarm from SongSparrows in Nature. Young hand-raised birds showed no alarm, while an adult male showed moderatealarm in Septemberand weak alarm the following April, but after experiencewith catsin the summerhe showedstrong alarm. 5. In Nature Sparrow Hawks (Falco sparverius) occasionally arousedalarm or fear, but were usually ignored. Other hawkspro- vokedfright. With the hand-raisedmale, stationarymodels of hawks evokedmoderate alarm, rapidly moving modelsfright. 6. A wild SongSparrow showed fear in the presenceof a mounted Barred Owl. Young hand-raisedbirds up to the age of three weeks gave no fear reaction to a mounted owl. Adult hand-raisedmales showedstrong alarm to a stationarymodel of an owl and fear to moving models. 7. Tests with cardboardmodels showed that the pattern of the owl is a totality in which the head and shadingare mostimportant. 8. Other mounted birds and cardboard models evoked moderate or weak alarm. 9. In Nature the Cowbird (Molothrusater) evokesstrong alarm from SongSparrows on their territories. It is questionableif this is basedon an inborn pattern. The hand-raisedmale SongSparrow, tested in autumn, showed only moderate and weak alarm to a mounted female Cowbird. 10. Snakesevoked curiosityin a hand-raisedSong Sparrowand also in a wild male. Small snakeswere treated as nest enemiesby a wild female. 11. Memory has been shownto be'of great importancein enemy recognition. In our experimentswe found that the memoryof cir- cumstancesconnected with strongalarm persisted after severalmonths (the piano without the owl). 12. 'Unpleasant'experience with an enemywill intensifythe reaction,while indifferentexperience will weakenthe reaction. Vol.sS] NICEAND TER PELKWYK, Enemy Recognition bySong Sparrow 2 ] 3

13. We think that owlsare recognized by SongSparrows in Nature largelythrough an inbornpattern, hawks through fast movements, and catsand possiblyCowbirds after conditioning.

REFERENCES ANTONIUS, O. 1939. Ueberdie $chlangenfurchtder Affen. Zeitschr.f. Tierpsych.,2: 295-296. Bo•z.•s, F. 1890, BarredOwls in captivity, Auk, 7: 101-114, BROCKNER,G. H. 1933. Untersuchungenzur Tierpsychologie,inbesondere zur Aufl6sungder Familie.Zeitschr. f.'Psych., 125: 1-110. GOETHE, F. 1937. Beobachtungenund Erfahrungenbei der Aufzuchtvon deutschemAuer- wild. DeutscheJagd, 1957,6 and 7. HANN, H. W. 1937. Life historyof the Oven-birdin southernMichigan. Wilson Bull., 49: 145-257. HEINROTH, O. AND M. 1924-32. V'6gel Mitteleuropas. Berlin. HERTZ, M. 1999. Beobachtungenan gefangenen Rabenv6geln. Psychol. Forsch., 8: $36-$97. K•c, W. N. 1931. A note on fear behaviorin youngrats, mice and birds. Joum. Comp. Psychol.,12: 117-121. K•mo, k 1940. Antipathy in the ScreechOwl. Auk, 57: 252-255. KPdiTZlC, H. 1939. Untersuchungenzur Biologie und Ethologie des Haselhuhns.Ber. Schlesischer Omith., 24: 1-25. 1940. Untersuchungenzur Lebensweisedes Moorschneehuhns(Lagopus l. lagopus)wShrend der Jugendentwicklung. Journ. f. Ornith.,88: 159--165. LOm•NZ, K. 1955. Der Kumpanin der Urnweltdes Vogels. Joum. f. Ornith., 85: 157-415. 1959. VergleichendeVerhaltensforschung. Zoologischer Anzeigcr, Supplement- band, 12: 69--102. N•CE, M. M. 1957. Studiesin the life history of the Song Sparrow. I. Trans. Linn. Soc. New York, 4: 1-247. RAND, A. L. 1937. Notes on the developmentof two youngBlue Jays (Cyanocittacristata). Proc. Linn. Soc. New York, no. 48: 27-59. SHERMAN, A. R. 1911. Nest life of the Sa'eech Owl. Auk, 28: 155-168. Sicx, H. 1959. Ueber die Dialectsbildungbeim "Regenfur" des Buchfinken. Journ. f. Ornith., 87: 568-592. STRAUSZ,E. 1958. Versuchean gcfangcnenRabenv6geln. Zcitschr. f. Ticrpsych.,2: 172--197. 2 l'i N,• At•i•TER PELKWYK, Enemy Recognition bySong Sparrow I-AukL April

THORNDIKE, F,,. L. 1899. The instinctivereactions of youngchicks. Psych.Rev., 6: 252-291. TINBF•RGEN•N. 1939. Why do birdsbehave as theydo? Bird-lore,41: 23-•0. 1940. "Die Ethologieals Hilfswissenschaftder Oekologie",Vortrag. Journ. f. Ornith., 88: 171-177. 5708 Kenwood Avenue Chicago,Illinois and Universityof Chicago Chicago,Illinois