BIRD-BANDING

A JOURNAL OF ORNITHOI•OGICAI, INVl•,STIGATION

VOL. 49, NO. 1 WINTER 1978 PAGES 1--100

NESTING BEHAVIOR OF THE GULL-BILLED TERN

BY HAROLD F. SEARS

INTRODUCTION The Gull-billed Tern (Gelochelidonnilotica) has a scattered distribution, and its breeding behavior has been little studied in the past (e.g., deWaard, 1952; Lind, 1963a; M•511er,1975). I have therefore made a relativelybroad studyof the breedingbehavior of thisspecies. I present my observationson nestingbehavior here and discussthe antipredator strategyof the Gull-billed Tern. The species'display behavior will be discussed elsewhere.

STUDY AREA AND METHODS I spent the springsand summersof 1971, 1972, and 1973 on an unnamed island 0.5 km south of the port at Morehead City, Carteret County, North Carolina.Beaufort Inlet liesto the southeast,and a Spar- tina marsh and Bogue Banks lie to the southwest. Early during eachbreeding season, I watchedthe ternsfrom exposed but distant positionsusing 8 x 24 binocularsor a 15-60 power zoom telescope.As colonieswere established,I positionedburlap blinds at their borders, and observations were made from within the blinds. I usedtime-lapse photography to studyincubation behavior. A Braun- Nizo S-80 movie camera was positionedon a tripod 20 to 30 m from the nest.It wasset to exposeone frame everyfour to five seconds(later determinedexactly) and wasleft in positionfor up to five hours. Gull-billedTerns are quite sensitiveto disturbance;therefore, I did not trap or mark any terns. This meant I could only rarely recognize individuals.The sexeswere not distinguishable,although Lind (1963a) thought he could distinguishmales by their heavier bill. No terns were naturally marked in any way, and only twice did I study pairs whose memberscould be distinguishedby tone of voice.I did band mostchicks with a Fish and Wildlife Serviceband, and on a few occasions,I placed a red dye in solution in a small mollusc shell in the bowl of the nest, thereby temporarily marking an incubatingadult. Such a mark faded after a few days.

RESULTS ColonySite and Composition The islandconsisted of recentlydredged spoil, which provided a va- riety of nestingsubstrates, including loose sand, sand mixed with shells, 2] H. F. Sears Bird-BandingWinter 1978 and firmer sandmixed with siltthat supportedlow vegetationof variable density.Gull-billed Terns in this area appearedto prefer a sitethat had a substrateof fairly rough texture (consistingof shells,sticks, and other debris as well as sand), that was sparselydotted with vegetation,and that was situatednear an area of relativelydense, low vegetation.The initial colonies of 1971 and 1972 were established in such sites. No Gull- billed Terns nestedin densevegetation, and nestingon bare substrate was associatedwith disturbedconditions. After widespreadnest deser- tion during unusuallycold wet weatherin 1972, presumedrenesting attemptswere made on virtually bare sand. Then, during the fall of 1972,the southwesternhalf of the islandwas largely buried under fresh spoil, and in 1973, one of the two major colonieswas establishedon bare substrate. Most of the colonies I studied contained both Common Terns (Sterna hirundo) and Black Skimmers (Rynchopsniger), as well as Gull-billed Terns. During 1971, the first colonyestablished contained at least 20 Gull-billed nests.Later, a secondand a third colonywere established, containingabout 9 and 3 Gull-billed nests,respectively. These were probably renesting attempts. During 1972 and 1973, two colonies formed initially (9 and 10 Gull-billed nestsin 1972 and 9 and 7 Gull- billed nestsin 1973), and again those that abandonedtheir nestsap- parently renestedin smallercolonies in other locations. In 1973,the averagedistance between a Gull-billednest and its near- estneighbor was about 10 m (Table 1). If only conspecificnearest neigh- bors are considered,the distancewas a little greater, 12 m. Of the 32 Gull-billed nestsconsidered, 16 had conspecificnearest neighbors, 10 had BlackSkimmer nearest neighbors, and 6 had CommonTern near- est neighbors.I collectedno quantitativedata on the total frequencies

TABLE l Averagedistance of Gull-billedTern nestsfrom their nearestneighbors in sevendifferent coloniesestablished during 1973•.

Neighbor Colony BlackSkimmer CommonTern Gull-billedTern

Far SW 2.7 (3) -- 7.0 (6) 5.6 W (14 May) 8.5 (2) 10.0 (3) 16.5 (2) 11.4 W (6 June) 2.0 (2) -- 4.0 (3) 3.2 Near SW 3.0 ( 1) -- -- 3.0 N 5.0 (2) -- 8.0 (1) 6.0 NW -- 5.0 (2) 10.0 (1) 6.7 S -- 42.0 (1) 29.3 (3) 32.5

• Distances in meters. Number of Gull-billed nests on which each mean is based is given in parenthesis. Vol.49, No. I NestingGull-billed Terns [3 of the three specieson the island,but the Gull-billedTern wascertainly in the minority,outnumbered by eachother speciesby at leasta factor of five.Thus, the Gull-billedTern tendsto nestnear conspecifics. TerritoryEstablishment Nestingterritories are establishedno more than a day or two before the first eggis laid. Establishmentof the territoryinvolves rapid local- izationof a pair'scourtship activity, at whichtime the nestis scraped out by both membersof the pair. The first eggswere laid about 5-7 May in 1971, 1972, and 1973. Nest Construction The Gull-billedTern tendsto nestnear an objectthat standsout from the substrate,such as a tuft of grass,a stick,or other pieceof flotsam. My studyarea was unusually bare, being composed largely of freshspoil, but nearlyhalf the nestsestablished during 1973were positioned near such a feature. The nestconsists of a depressionin the sand,a surroundingrim of sand raised above the level of the substrate, and an accumulation of shells,sticks, or grassblades on and around the rim. The exactnature of the nestvaries considerably from nestto nestand f¾omday to day, largely dependingupon the area in which the nest is located.At a nest on an open shellyarea, many shellsare incorporatedinto the nestand nestrim. At a neston a shrubby,grassy area, twigsand grassesare used. In open areas,nests become well defined and complexduring periods of calm weatherbut may be obliteratedduring heavywind or rain. In protectedareas, nests remain deep and the rims conspicuous. The constructionof the nestinvolves three relativelydistinct behavior patterns:scraping, sideways-throwing, and sideways-building.Nest ma- terial is not broughtto the nest from a distance.A tern beginsnest constructionby scrapingout a shallowdepression in the sand.The tern kicksits feet alternatelywhile restingon its breastand so pushessand backwardsand onto the nestrim. Subsequentdeepening of this bowl and later clearingit of blown sandinvolve the samemovements. During sideways-throwing,shells or other items are tossedfrom the immediate vicinity of the nest toward the nest. The tern standswith wingsslightly abducted and drooped. The tern stretchesits neck out horizontally,reaches down to the substrate,grasps the item in its bill, and tossesit to one sideand backwardwith a quickand relativelyster- eotyped turn of the head. The item is tossedas far as a meter to the rear of the bird. Sideways-throwingis occasionallyperformed by an incubatingtern thatrises from the nestand walksaway, sideways-throw- •ng repeatedly. The tern then walks back to the nest and resumesin- cubation.Sideways-throwing isalso commonly performed during or im- mediatelyfollowing the nestexchange. With the repeatedperformance of thisbehavior through the incubationperiod, material is slowly moved from the surroundingarea toward the nest. 4] H. F. Sears Bird-BandingWinter 1978

Sideways-buildingis a more variableand more well-directedbehavior wherebymaterial is actuallyplaced onto the nest rim. It is performed by incubatingterns or by ternsstanding near the nestbefore beginning to incubate. Although the sideways-throwis relatively stereotypedin form, it is frequently oriented with respectto the nest,just as sideways- building is. Perhapssideways-throwing and sideways-buildingare two ends of a continuum. At a distance from the nest, the movement is relatively stereotypedand undirected, but the closerthe tern is to the nest, the more deliberate and well-directed the movement becomes (Harrison, 1967).

Clutch Size The Gull-billedTern producesclutches of one to three eggs(• = 2.0, n = 41). Most of the pairs that establishednests at the beginningof the 1973 season(before 15 May) produced clutchesof two or three (• = 2.5, n = 16). Most of thosethat nestedlater producedclutches of one or two (• = 1.7, n = 25). A decreasein the average clutch size during the nesting seasonis a trend found in many birds (Bergman, 1953; Cullen, 1956; Langham, 1968, 1974). However, these data probably give a biased picture of clutch size in thisspecies because some clutches could have been reduced by predation before I discoveredthem or could have been abandonedbefore laying was completed.Pemberton (1927) describesone nest containingfour eggs.Bent (1963) reportsthat the clutchsize is two or three, rarely four. The eggswere laid at a rate of lessthan one per day. Five three-egg clutcheswere laid over the courseof at least four days,and three were laid over the courseof at leastfive days. IncubationSpell An incubationor sitting spell is an uninterrupted period of incuba- tion, separatedfrom the precedingand following spellsby suchoff-the- nest behavior as nest exchanges,sideways-throwing, defecation, or alarm flight. Using time-lapsephotography, I monitoredincubation behavior for two to five hours on each of 36 separateoccasions at 12 different nests. The duration of the incubationspell ranged from lessthan 1 minute to over 194 minutes,and the daily mean duration varied almostas much' 5 minutes to 152 minutes (• = 42, SD = 30, n = 36). To explain someof this variability,I measuredthe degree of associ- ation betweenthe daily mean duration of the incubationspell and each of two temporal parameters,date and days since completionof the clutch,and eight weatherparameters: daily rainfall, high temperature, and low temperature(taken from ClimatologicalData for MoreheadCity publishedby the National Oceanicand AtmosphericAdministration) and skycover, windspeed, wetbulb temperature, dewpoint, and relative humidity (taken from LocalClimatological Data for Cape Hatteras, N.C. also publishedby NOAA). I calculatedSpearman's Rank Correlation Vol.4o, No. 1 Ne,stingGull-billed Terns [5

160

120

80 o 40 o o o x

õ o o o o o o

I 5 10 15 20 Stage of Incubation (days) FIGURE1. The changein the mean durationof the incubationspell through the incu- bationperiod (day 1 = first day of incubationof full clutch).The leastsquares regres- sion line (y = 1.70x + 27.66) was calculatedfrom the data for unhatchedclutches only (o = unhatchedclutch; x = one or more eggshatched).

Coefficientfor each pair of variablesand tested Ho: rs = 0 with a one- tailed t-test (Siegel,1956). Spell duration is not significantlycorrelated with any of the weather parameters,nor with the temporal parameters when the entire nestingperiod is consideredin the analysis.However, if only the first three weeksof the nestingperiod are considered,spell duration is significantlycorrelated with the stage of incubation (rs = 0.48) and thereforewith the date (rs = 0.37). Figure 1 illustratesthis associationin more detail. Spell duration tends to increasethrough the 6] H. F. Sears Bird-BandingWinter 1978 incubation period, peaks around the time of hatching, and falls sharply thereafter. The interval separatingtwo incubationspells normally rangesfrom a few seconds to a few minutes. At the 12 nests discussed above, the relative amount of time a nest wasleft uncoveredranged up to 8% (• = 1.7%, SD = 1.6, n = 36). During the first three weeksof incubation, this variable is significantlynegatively correlated with the stageof in- cubation(rs = -0.52) and with the date (rs = -0.39). During the entire nestingseason, it is significantlynegatively correlated with daily rainfall (rs = -0.38) and with the high and low temperaturesfor the day (rs = -0.41 and -0.38, respectively). It has often been found that birds becomeincreasingly attentive at the nestas the incubationperiod progressesand during periodsof harsh weather (Cullen, 1956; Beer, 1961; Howell and Bartholomew, 1962; Baerends and Drent, 1970; Drent, 1973). Defecation and Regurgitation Adults apparently do not defecatein or near the nest. Fecesdo not accumulate around the nest, and on several occasions, I saw an incu- bating tern walk or fly from the nest, defecate,and return. It is also common for a tern, after a nest exchange,to defecateas it walksfrom the nest. Adults regurgitatepellets of undigestedmaterial onto the nest,how- ever. I collectedmany pelletsfrom nests,nest rims, and nestvicinities. At some nests,I collecteda pellet daily for severalconsecutive days.

Egg-retrieval The Gull-billedTern retrieveseggs that havebecome displaced from the nest, but it doesnot do so consistently.On five different occasions, I movedan egg to the rim of a nestor just beyondthe rim, leavingone or two eggsremaining in the nest.Three of theseexperiments resulted in retrieval, but only after 10 to 15 minutesof incubationon the nest, and two experimentswere terminatedafter 30 minutesof incubation without retrieval. On the other hand, I have seen the retrieval of an eggshell,which had been movedto the nestrim during a nestexchange, and the retrievalof a pieceof white plasticcup that I had placedon the nest rim. Both of these retrievals occurred soon after the tern settled onto the nest. During retrieval,the eggis drawn under the bodywith the bill. The mostdistant retrieval was performed in two stages,the tern resettling on the nest between them. Distant retrieval in stagesseems to be a commontechnique in terns (Marshall, 1943; Hawksley,1950; K/Srner, 1966).

Reactions to Predators No evidenceof predation on adult Gull-billed Terns was seen,but eggsand youngwere obviouslyvulnerable. Humans and dogsoccasion- Vol.49, •o. I NestingGull-billed Terns [7

ally came onto the island, and I once saw a dog running along the periphery of a colonycarrying a juvenile tern in its mouth. A few rats (probablyRattus norvegicus) had burrowson the island,and I once fol- lowed rat tracksinto a colonyand past an empty Common Tern nest that had held eggsthe previousday. OccasionallyI saw trespassing chicksseverely pecked by neighboringGull-billed and Common terns, and both eggsand chickswere taken by Laughing Gulls (Larusatricilla) and by immature Herring Gulls (L. argentatus).In addition, I have seen adultsreact defensivelytoward the Ruddy Turnstone (Arenariainterpres) (in contrast,Wilson's Plovers, Charadrius wilsonia, were ignored) and the Red-wingedBlackbird (Agelaius phoeniceus), which have been known to take tern eggs(Pessino, 1968; Crossinand Huber, 1970; Parkeset al., 1971; Dinsmore,1972), and alsotoward the SnowyEgret (Egrettathula) and EasternMeadowlark (Sturnella magna). During 1973, nest success was 40% (17 broods/43 nests). Most gulls flying near the colony elicit no reaction from the terns. I have alsoseen an adult simplylead its young away from a nearbygull, and in another case,fly closerto its young. An approachingaerial pred- ator elicits aggressivepostures and vocalizations.One or more terns then fly up, chasethe predator,and continueto vocalize.An approach- ing ground predator elicitsthe same aggressivedisplays, and at least severalindividuals finally fly up. They hover over the predatorand, one by one, dive toward it and then swoopback up to the flock. Callsare often uttered throughoutthe dive. At the lowestpoint, the tern often calls,defecates, and strikesthe predator with its bill or feet. The responsivenessof the pair changesthrough the incubationpe- riod. Incubatingterns are particularlyreluctant to leavethe nestaround the time of hatching. I have approachedor entered the colony and noticed that adults with hatching eggsor eggs about to hatch did not leave their nests, whereas those at other nests did; or that adults with hatchingeggs were quicklyback on their nestsafter I entered the blind, whereas others were much slower to return. At the same time, those adults that do respond to intruders do so with increasedintensity late in incubation. The effect of group attackis usuallyto repel the potentialpredator. I have seen picnickersapproach a colony but turn abruptly and walk backto the shoreafter a smallflock of terns begandiving at them. The dog mentionedabove was harried by a smallflock of terns,and it loped along through the thick vegetationbordering the colonyrather than through the open area occupiedby the colony itself. Diving terns also occasionallypunctured my fiber-boardhelmet with their bills;a peck of this intensityis probablya real deterrent to potential predators. EggshellRemoval Gull-billedTerns seemto haveonly a weaktendency to removeegg- shells.I have seen only two natural casesof eggshellremoval. In one case,the tern walked from the nest, carrying the shell,and dropped it two or three meters from the nest. The chick was still wet. In the second 8] H. F. Sears Bird-BandingWinter 1978 case,the adult flew from the nestcalling and droppedthe shell30 to 40 metersaway. In contrast,I havewatched several nests at whichthe shells were definitely not removed. Shellsremained in two nestsand within a meter of sevennests for one day after hatching,in one nestand within a meter of three nestsfor two daysafter hatching,and within a meter of six nestsuntil after the parentsand chicksleft the nest. I introduceda varietyof objectsinto variousGull-billed nests, includ- ing tern eggshells,Willet (Catoptrophorussemipalmatus) eggshells, white chickeneggshells, a red and brassshotgun shell, and a pieceof white plasticdrinking cup (about 6 cm in longestdimension). Of the 15 tern and Willet eggshellsintroduced, seven were removedwithin the fbur to fivehour observationperiod and sevenwere not. At the remainingnest, the pair and their chicksleft the nest.Of the sevenremovals, only one consistedof a long-distanceaerial carry. The others were removalsto the nestrim or only slightlybeyond. Of the 12 chickeneggshells introduced, seven were carriedaway and five were not, and the shotgunshell was not moved,even after two days. The pieceof plasticlay on the nestrim as the tern settledonto the nest, and the tern immediatelyretrieved it as if it were an intact egg. Later, the tern's mate arrived to assume incubation. This individual dove at the plasticas if it were a predator and, on the third dive, grabbedthe plastic,flew up and away,and droppedit 10 to 20 metersfrom the nest. Alarm or disturbancewas evident after someof the other experimen- tal introductions.The tern approachingthe shotgunshell in the nest assumedan alarm posture, called, and walked back and forth before the nest before finally settlingonto it. This reactionor a lessintense form of it wasalso elicited by someof the chickeneggshells. The tern and Willet eggshellsdid not tend to elicit suchalarm behavior. Thus, the Gull-billedTern tendsto removeconspicuous objects from the nest, but suchobjects are not alwaysremoved, and they are often movedonly a shortdistance. My few experimentsrevealed no increased tendencyto removeobjects late in the incubationperiod. The wind playsa fairly important part in what removaldoes occur. I sawa tern apparentlyunintentionally nudge an eggshellto the nest rim during a nestexchange. The wind then blew the shellaway. Leavingthe Nest The adultsand chicksremain at the nestfor a few daysat most (see below).After this time, the family leavesthe nestsite for a more shel- tered area. During 1972 and 1973, I watched14 different familiesfor varyingperiods after this departure. Usuallythe adultsseem to direct the movementfrom the nest.They walk from the nestcalling repeatedly and occasionallypausing to scrape, and the chicksfollow, usually vocalizing. However, the chickssometimes direct movementaway from the nest.At one nest,there wasa two-day- old chickand an eggthat hatchedlater that day. I had removedthe egg and chickfor experimentalreasons; upon replacing them, an adultland- Vol.49, No. I NestingGull-billed Terns [9 ed, and the chick immediately walked h'om the nest and continued severalmeters past the adult. The adult overtook the chick and led it a distance of 45 meters. On another occasion, an adult landed at a nest, called, and led the chick from the nest. After a few meters, the adult stoppedand began to cover the chick, but the chick walked on. The adult roseand followed,calling. The chickfinally stopped, and the adult covered it. In a third case, an adult led a chick a meter or so and con- tinued to walk and call, but the chickhad stopped,and the adult turned and came back to the chick. Three main factorsapparently control the timing, distance,and di- rectionof movementaway from the nest:ages of the chicks,disturbance to the family, and distributionof vegetationaround the nest. The age of the youngestchick in the brood exerciseslittle controlover the timing of departure from the nest. Nests are left as early as one day after hatching,and the intervalfrom the hatchingof the firstegg to departure is about the same (about 2.5 days) for clutchesof one (n = 4) or two (n = 8), and apparentlyfor thoseof three as well (n = 2). However, the presenceof youngchicks seems to affect the distancetraveled. During the firsthour after leavingthe nest,broods with one-day-oldchicks were led an averageof 26 meters (n = 6) and broodswith only older chicks were led an averageof 80 meters(n = 5). Disturbanceto the adultsor chicksoften appearsto initiate departure from the nest. I witnessedeight of the 18 departuresthat occurred during 1972 and 1973, and sevenwere initiated within a few minutes of my arrival. Similarly,my arrival seemedto trigger many instancesof post-departuremovement. I witnessed19 movementsthat occurredat leastone day after departure from the nest. Of all 27 observedmove- ments, 19 were initiatedwithin 15 minutesof my arrival, and only eight were initiated later. This distribution is significantlydifferent from a 50:50 distribution,which is a conservativenull hypothesissince all ob- servationperiods exceeded 30 minutes(X 2 = 4.48, 1 d.f., P • 0.05). After moving away from the nest, the family tends to settle among relativelythick vegetation.I classifiedeach stopping point during move- ment following departure from the nest accordingto presenceor ab- senceof vegetationand accordingto the length of stayat the site (Table 2). Areas of vegetationwere occupiedabout as often as bare sites.Since at least80% of the nestingarea was bare, thisalone indicates a preference for shelteredsites. The tablealso shows that staysin areasof vegetation tend to be longerthan thoseon bareareas (X 2 = 19.9, 1 d.f., P • 0.01). The youngare ultimatelyled to the beachesand mudflatsthat make up the shores of the island. I quickly lost contact with some of the families after departure from the nest, but even so, six reached the water, sevenreached the sanddike that partiallysurrounded the island not far abovehigh water, and only five were still on the islandinterior at the time I last saw them. In all, most families traveled at least a few hundred metersbefore the young flew. One family movedabout a ki- lometer. H. F. Sears Bird-Banding 10] Winter 1978

T^BLE 2 Relationshipbetween length of stayand presenceor absenceof vegetation.

Length of stay <2 hr >2 hr Total

Vegetationpresent 2 23 25 N one 16 13 29

DISCUSSION Featuresthat have the effect of reducing predation on eggsand young (seeCullen, 1960;Tinbergen, 1963 and 1967; and Lack, 1967 and 1968 for pertinentgeneral discussions) can be classifiedas thosethat tend to keep predatorsfrom the nestingarea, thosethat tend to hide the clutch or brood from predatorsin the area, or thosethat tend to repel pred- ators. A feature of Gull-billednesting behavior that tendsto keep predators from the nestingarea is the isolatedposition of the colony.The island site of the coloniesI studied and the surrounding Coast Guard and State Park property greatly reduced land accessto the ternery by man and dog, probablythe two most important terrestrial predatorsin the area. Isolation of the colonyis much more stronglydeveloped as an anti-predatoradaptation in other terns (e.g.Sterna maximus and S. sand- vicensisamong temperate species (Lind, 1963b;Kale et al., 1965;Buckley and Buckley, 1972)). Many more aspectsof the Gull-billedTern's nestingbehavior tend to prevent a predator from finding the clutch or brood, oncethe predator does enter the breeding area. The nestsin the colony are relatively dispersed,and severalsmall colonies tend to form rather than one large one. It wasclear that the disturbancecaused by my activitieswas partially responsiblefor this nest distribution.For example, during 1972, my presenceappeared to prevent the establishmentof a colonyin what had been the preferred area in 1971. When I moved my blind to the area that wasbecoming occupied, further nestingceased and nestingbegan in a still lesssuitable area. My activityin the first colonyto form in 1973 triggeredthe establishmentof the secondmain colonyof that year,and my subsequentdiscovery and censusof the secondcolony contrib- uted to the return of some individuals to the first. The relocation of renestingattempts might alsohave been due to my presence. However, the nest distributionI observedwas not grosslyabnormal. The mean nearest-neighbordistance that I measured (about 10 m) is much greater than the 15 inches(38 cm) to 3 feet (91.5 cm) reported by Cain (1933) and the 1.5 m reportedby Lind (1963a),but it is similar to the 20 feet (6.1 m) reported by Pemberton(1927), and it may reflect Vol.49. No. 1 NestingGull-billed Terns [ 11 either the similarlyopen nature of the two sitesor the similarlylarge area available.Also, although coloniesof 100, 200, or even 300 nests havebeen reported (Glegg, 1931; Buckley,pers. comm.; Dement'ev and Gladkov, 1969), the typical colonyapparently contains about 10 to 20 nests (Bent, 1963; Jensen, 1946; Stewart and Robbins, 1947; Lind, 1963a),as many of my coloniesdid. Thus, nestdispersion and the sen- sitivityto disturbancethat contributesto it are probablyadaptations that increasethe difficulty the predator has in locatingnests. The tendencyto form mixed coloniesmight contributeto the con- cealmentof the nestand contents.Croze (1970) hassuggested that the SandwichTern (Sternasandvicensis) derives protection of this sort by its habit of nestingwith other speciesof terns and gulls.The Sandwich Tern is unusualin that it defecatesaround the nest,producing a whitish backgroundfor unusuallylight eggs. Predatorsdevelop a specific searchingimage (Tinbergen, 1960) for the more commonnests of the specieswith which the SandwichTern nestsand thus overlookthose of the SandwichTern itself. Analogously,the Gull-billed Tern (and the Common Tern) tends to accumulate debris about its nest, in which it lays relativelydark eggs,and the Black Skimmersin the colonyaccu- mulate no material and lay relativelylight eggs.The nestsand eggsof the different speciesare alsoof different sizes,and thisvariation within the colony must make it more difficult for a predator to develop a definite specificsearching image. The tendencyto nest near a plant or other objectmight alsohelp to hide the nest, sincesuch a site is relatively visuallycomplex. It might alsotend to prevent birds of prey from getting a cleanswoop at the nest (Munro, 1960). Adult Gull-billed Terns do not defecate near the nest, and the com- parisonbetween this inhibition and the lackof inhibitionassociated with the regurgitationof pelletsindicates that camouflage,rather than san- itation, underlies the inhibition of defecation. The feces are largely white and if localizedthickly about the nest would becomeobvious to a predator (Cullen, 1960). The regurgitatedpellets are dark, mottled collectionsof vertebratebones and insectexoskeletons, and they disin- tegrate quickly to an inconspicuousscattering among the debrisabout the nest. Eggshellsnear the nest can alsojeopardize the brood through their conspicuousnessto predators (Tinbergen et al., 1962);thus the tendency of the Gull-billed Tern to remove them is alsoan adaptationthat serves to camouflagethe nest.However, this tendency is apparently quite weak. Most terns seem to remove the eggshellssoon after hatching (Jones, 1906;Marpies and Marpies,1934; Palmer, 1941; Hawksley, 1950; Cuth- bert, 1954; Goodwin, 1960; Sch/3nert, 1961), and at least the Common Tern removesother conspicuousobjects as well (Pettingill,1939; Pal- mer, 1941), althoughthe consistencywith which this is done is not usu- ally stated.Cullen (1956) noted that Arctic Terns (Sternaparadisaea) do not alwaysremove eggshells, although they usually do; Dinsmore(1972) 12] H. F. Sears Bird-BandingWinter 1978

reported that SootyTerns (Sternafuscata)seldom do, and Buckleyand Buckley(1972) reportedthat RoyalTerns never removethe eggshells. The weak expressionof this adaptationin the Gull-billed Tern could be related to the fact that they, like the RoyalTerns (Buckleyand Buck- ley, 1972), quicklyleave the nest after the eggshatch. Predationon chicksin the nestmight be relativelyslight. Most of my observationsof predationwere madeduring the abnormallybad weatherof 1972,when many pairs abandonedtheir nestsand later renested.Langham (1968, 1972) has shownpredation to be an insignificantcomponent of chick mortality in the Common, Arctic, Roseate(Sterna dougallii), and Sand- wich terns that he studied and that starvation is the most important causeof mortality during the first week after hatching.In contrast, predation does seemto be significantin the Black-headedGull (Larus ridibundus)(Tinbergen et al., 1962). Thus, many factorsassociated with the nest and its locationappear to hide the nestfrom predators.The opennessof the colonysite allows the incubatingterns a clear view of their surroundingsand the chanceto detect an approachingpredator while it is still a significantdistance away.The existenceof many observersin a colonyfurther increasesthe probabilityof earlypredator detection (Siegfried and Underhill, 1975). The conspicuousadult can then fly and thus permit the crypticfeatures of the nest to have their effect. However, predationis probablynot the major selectiveagent responsiblefor the presentform of the nest.Most nestsappear rather conspicuous,and some are quite substantialand stand out clearly againsttheir sand and shell background.The nest is probablyalso important in protectingthe clutch and brood from other stresses, such as wind-blown sand. Two other aspectsof nestingincrease the difficultywith whicha pred- ator locates the nest or brood. The ease with which the terns relocate nestingattempts, following early failure, might be an adaptationserving to removethe pair from a disturbedarea (Buckleyand Buckley,1972), and the tendencyto leave the nest quickly after hatchingcertainly has this effect. Most ground-nestingterns do leave the nest after hatching,but the distancetraveled varies considerably. The Arctic Tern usuallyremains within a few meters of the nest until fledging, even at nestson open sand (Cullen, 1956). The young apparentlyremain on the nestingter- ritory and even defend it (Bullough, 1942; Hawksley, 1950). Cullen noted that adultswill occasionallytry to lead their youngas far as 10 to 20 metersfrom the nestand that they sometimessucceed, but usually, either the adultsor the youngstop and return to the territory.Similarly, the Sooty Tern abandonsthe nest but remains near the territory and returns to the territory eachnight until fledging(Dinsmore, 1972). The marsh terns (Black and Whiskered) (Chlidoniasniger and C. hy- bridus)seem to be somewhatless tied to the nest. They leave the nest and swimabout in the nearbyvegetation only a day or soafter hatching, but they remain near the nest for one or two weeks, and the adults sometimeslure them backto the nestduring this time (Hoffmann, 1926; Vol.49, •o. 1 NestingGull-billed Terns [ 13

Baggermanet al., 1956; Swift, 1960; Fuggles-Couchman,1962; De- ment'evand Gladkov, 1969). After this time, the young wander farther until fledging.Goodwin (1960) suggeststhat nest-areaabandonment can be more rapid in the Black Tern and that early abandonmentis asso- ciated with unusual disturbance of the nest. In contrastto thesespecies are the LeastTern (Moseley,pers. comm.), the SandwichTern (Bickerton, 1912; Dircksen, 1932;J. van den Assem, Ms), the RoyalTern (Buckleyand Buckley,1972), and the Gull-billed Tern (M•ller, 1975; this study),all of which leave the nest soonafter hatchingand wanderrelatively far from it. Much of this variationcan probablybe explainedby the amount of disturbanceto which the terns were subjectedand the amount of shelter the surroundingvegetation provided. Finally,the mostconspicuous anti-predator adaptations are thosethat repel predatorswhich have found the nest or brood or are about to find it. These adaptationsare individualthreat and attack,the increased tendencyto attacklate in incubationand during hatching,group attack, and the colonial nestingand synchronouslaying that permits group attack (Kruuk, 1964; Patterson, 1965; Sears,Ms; but seeLemmetyinen, 1971 for counter-argument).

SUMMARY Nestingbehavior of the Gull-billedTern wasstudied for three seasons in smallmixed-species colonies on an islandjust off the coastof North Carolina. Many aspectsof the species'parental behavior appear to pro- tect the clutchor brood from predators.These include:isolated location of the colony,degree of nestdispersion, tendency to form mixed-species colonies,tendency to nest near a plant or other object,tendency not to defecate near the nest, tendency to remove eggshellsand other con- spicuousobjects from the nest, opennessof the colonysite, easewith which the terns relocatenesting attempts, tendency to leave the nest soon after hatching,threat and attack behavior, group attack, and the colonialnesting and synchronouslaying that permit group attack. Other aspectsof the Gull-billedTern's parentalbehavior seem to con- tribute to the productionof the pair in other ways.During periodsof rain or high temperature,adults cover the nesta greaterpercentage of the time. Average clutch size decreasesthrough the breeding season; this probablycompensates for the inabilityof late nestersto raisea large brood. Finally, in many respects,adults put increasingeffort into in- cubationas the incubationperiod progresses. Several different processes might explainthis trend, but its adaptivevalue is clear:the clutchbe- comesincreasingly hard to replacethrough the incubationperiod, and at the time of hatchingit is maximallyvulnerable.

ACKNOWLEDGMENTS Many people have assistedme in this studyby providingequipment, supplies,or ideas,and I am gratefulto eachof them. I particularlywant 14] H. F. Sears Bird-SandingWinter 1978 to thank Drs. H. C. Mueller and R. H. Wiley of the Universityof North Carolinaat Chapel Hill, Dr. A. F. Chestnutof the U.N.C. Institute of Marine Sciences,the Frank M. Chapman Memorial Fund of the Amer- ican Museum of Natural History, the H. van PetersWilson Fund of U.N.C., and my wife, M. L. Sears.In addition, I am grateful to an anonymousreviewer who helpedimprove my presentation.

LITERATURE CITED

BAERENr•S,G. P., ANr•R. H. DRENX. 1970. The Herring Gull and its egg. Behaviour (suppl.), 17: 1-312. BAGGERMAN,B., g. P. BAERENDS,H. S. HEIKENS,ANDJ. H. MOOK.1956. Observationson the behaviourof the BlackTern, Chlidoniasn. niger(L.), in the breedingarea. Ardea, 44:1-71. BEER,C. G. 1961. Incubationand nest building behaviour of Black-headedGulls. I: Incubationbehaviour in the incubationperiod. Behaviour, 18: 62-106. BENX,A. C. 1963. Life Histories of North American Gulls and Terns. New York, Dover. BERGMAN,G. 1953. Verhaltenund Biologieder Raubseeschwalbe(Hydroprogne tschegrava). Acta Zool. Fenn., 77: 1-50. BICKERTON,W. 1912. The Home-life of the Terns or Sea Swallows.London, Witherby. BUCKrE¾,F. G., ANr•P. A. BUCKrE¾.1972. The breedingecology of Royal Terns Sterna (Thalasseus) maxima maxima. Ibis, 114: 344-359. BULLOUGH, W. S. 1942. Observations on the colonies of the Arctic Tern on the Fame Islands. Proc. Zool. Soc. Lond., 112A: 1-12. CAIN,W. 1933. Nestingof the Gull-billedTern. SouthAustralian Ornithologist, 12: 28-30. CROSSIN,R. S., ANr• L. N. HyBER.1970. Sooty Tern eggpredation by Ruddy Turnstones. Condor, 72: 372-373. CROZE,H. 1970. Searchingimage in Carrion Crows.Z. Tierpsychol.(suppl.), 5: 1-85. CUrtEN,J. M. 1956. A studyof thebehaviour of theArctic Tern (Sternamacrura). Unpubl. doctoral dissertation,Oxford, Wodham College. ß 1960. Some adaptationsin the nestingbehaviour of terns.Proc. 12th Intern. Ornithol.Congr., 153-157. CUTHBERT,N. L. 1954. A nestingstudy of the BlackTern in Michigan.Auk, 71: 36-63. DEMENT'EV,G. P., ANDN. A. GLADKOV(eds.). 1969. Birds of the SovietUnion. Jerusalem, Israel Program for ScientificTranslations. DINSMORE,J.J. 1972. SootyTern behavior.Bull. Fla. StateMus., Biol. Sci., 16: 129-179. DIRCKSEN,R. 1932. Die Biologiedes Austernfischers,der Brandseeschwalbeund der Kfistenseeschwalbenach Beobachtungenund Untersuchungenauf Norderoog.J. Ornithol. , 80:427-521. DRENT,R. 1973. The natural historyof incubation,in BreedingBiology of Birds,D. S. Farner, ed. p. 262-311. Washington,Nat. Acad. Sci. FUGGrEs-CoucHMAN,P. 1962. Nestingof the WhiskeredTern in Tanganyika.Ibis, 104: 563-564. GREGG,W. E. 1931. The birds of "L'Ile de la Camargueet la Petite Camargue."Ibis (ser. 13), 1: 209-241,419•446. GOODWIN,R. E. 1960. A studyof the ethologyof the BlackTern, Chlidoniasniger suri- namensis(Gmelin). Unpubl. doctoral dissertation, Ithaca, N.Y., CornellUniversity. HARRISON,C. J. O. 1967. Sideways-throwingand sideways-buildingin birds.Ibis, 109: 539-551. HAWKSrE¾,O. 1950. A studyof the behaviorand ecologyof the Arctic Tern, Sterna paradisaeaBrunnich. Unpubl. doctoral dissertation, Ithaca, N.Y., CornellUniversity. HOFFMANN,P. W. 1926. Nestingof the BlackTern in Wisconsin.Auk, 43: 86-87. HO%•ELL,T. R., ANDG. A. BARTHOLOMEWß1962. Temperature regulation in the Sooty Tern. Ibis, 104: 98-105. Vol.49, No. 1 NestingGull-billed Terns [ 15

JE•sE•, P. V. 1946. Nogle Iagttogelserover Sandternens(Gelochelidon nilotica Gm.) Bio- 1ogi.Dansk Ornithologist Forening Tidsskri[t, 4(•: 80-96. Jo•Es, L. 1906. A contributionto the life historyof Commonand RoseateTerns. Wilson Bull., 18.' 35-47. K^L•, H. W., G. W. ScIPL•, ^•D I.R. TOMKIN$. 1965. The Royal Tern colonyof Little Egg Island, Georgia.Bird-Banding, 36: 21-27. K'•R•R, H. K. 1966.Zur Eirollbewegungder K6stenseeschwalbe (Sterna macrura Naum.). Z. Tierpsychol.,23: 315-323. KRVVK,H. 1964. Predatorsand anti-predatorbehavior of the Black-headedGull. Behav- iour (suppl.), 11: 1-129. L^cK, D. 1967. Interrelationshipsin breeding adaptationsas shownby marine birds. Proc. 14th Intern. Ornithol.Congr., 3-42. ß 1968. EcologicalAdaptations for Breedingin Birds.I•ondon, Chapman and Hall. L^•G•^M, N. P. E. 1968. The comparativebiology of terns,Sterna spp. Unpubl. doctoral dissertation,Grey College,University of Durham. 1972. Chick survivalin terns (Sternaspp.) with particular referenceto the Com- mon Tern. J. Anim. Ecol.,41: 385-395. ß 1974. Comparativebreeding biology of the SandwichTern. Auk, 91: 255-277. LEMMETYINEN, R. 1971. Nest defense behavior of Common and Arctic Terns and its effect on the successachieved by predators.Ornis Fenn., 48: 13-24. LIND, H. 1963a. The reproductivebehaviour of the Gull-billed Tern, Sternanilotica Gme- lin. VidenskabeligeMeddelese, 125: 407-448. 1963b. (Noteson socialbehavior in terns).Dansk Ornithologisk Forening Tidsskrift, 57'.'155-175. M^RP•S, G., ^•D A. M^Re•ES. 1934. Sea Terns or Sea Swallows.London, Country Life. M^Rs•^•, N. 1943. Factors in the incubation behavior of the Common Tern. Auk, 6(•: 574-588. M•L•R, A. P. 1975. (The breedingbiology of the Gull-billedTern Gelochelidonn. nilotica Gm. in Denmark). DanskOrnithologisk Forening Tidsskrift, 69: 9-18. Mv•Ro, G. C. 1960. Birds of Hawaii. Vermont, BridgewayPress. P^•M•R, R. S. 1941. A behavior study of the Common Tern. Proc.Boston Soc. Natural History,42:1-119. P^R•S, K. C., A. PooL•, ^•D H. L^e•^•. 1971. The Ruddy Turnstoneas an egg pred- ator. Wilson Bull., 83: 306-308. P^•r•rERso•,I.J. 1965. Timing and spacingof broodsin the Black-headedGull. Ibis, 1(•7: 433-459. P•R•rO•, J. R. 1927. The AmericanGull-billed Tern breedingin California.Condor, 29.' 253-258. PESSINO,C. M. 1968. Red-winged Blackbird destroyseggs of Common and Roseate Terns. Auk, 85: 513. PET.T.INGILL, O. S., JR. 1939. Historyof 100 nestsof the ArcticTern. Auk, 56: 420-428. SC•O•R•, C. 1961. Zur Brutbiologieund Ethologieder Zwergseeschwalbe(Sterna a. alb![ronsPallas), in Beitr•igezur KenntnisDeutscher V6gel, H. Schildmacher,ed. p. 131-187. Jena, GustavFischer. Si•GE•, S. 1956. NonparametricStatistics for the BehavioralSciences. New York, Mc- Graw-Hill. SIEGFRIED,W. R., ANDL. g. UNDERHILL.1975. Flockingas an anti-predatorstrategy in doves. Anim. Behav., 23: 504-508. Sq71•W^RT,R. E., ANDC. S. ROI313IN$.1947. Recentobservations on Maryland birds.Auk, 64: 266-274. Swim'r,J.j. 1960. Notes on the behaviour of Whiskered Terns. Brit. Birds, 53.' 559-572. TINBERGEN,L. 1960. The natural control of insectsin pinewoods,I. Factorsinfluencing the intensityof predationby songbirds.Archives Neerlandaises deZoologie, 13• 265-343. TINBERGEN,N. 1963. On adaptiveradiation in gulls(tribe Larini). ZoologischeMededelingen Ri]ksmuseumvan Natuurli]keHistorie Leiden, 39.' 209-223. 16] H. F. Sears Bird-Banding Winter 1978

ß 1967. Adaptivefeatures of the Black-headedGull Larusridibundus L. Proc.14th Intern. Ornithol.Congr., 43-59. TINBERGEN,N., G. J. BROEKHUYSEN,F. FEEKES, J. C. W. HOUGHTON,H. KRUUK,AND E. SzuI•c. 1962. Egg shell removal by the Black-headedGull, Larus ridibundusL.; a behaviourcomponent of camouflage.Behaviour, 19:74-117. DEW^^RD, S. 1952. On the Gull-billed Terns at "de Beer" near Hook-of-Holland in 1949. Brit. Birds, 45: 339-341. Departmentof Zoology,University of NorthCarolina, Chapel Hill, NC 27514. (Presentaddress: University of SouthCarolina, Union, SC 29379.) Received 23 December1976, accepted17 August 1977.