Ecology of a Breeding Population of Arctic Terns

BIRD-BANDING

A JOURNAL OF ORNITHOLOGICAL INVESTIGATION

Vol. XXVIII April, 1957 No. 2

ECOLOGY OF A BREEDING POPULATION OF ARCTIC TERNS

By OSCARHAWKSLEY INTRODUCTION Biologiststoday are •nuch concernedwith the ecologicalfactors which controlnatural populations. Avian populations,particularly of colonialspecies, lend the•nselveswell to investigationsof the survival and mortalityrates of offspringup to the ageof attainingflight. Factors whichcontrol the successof the socialunit or colonyas a wholeduring the reproductiveperiod may alsobe readilystudied. The presentstudy wasmade during the breedingseasons of 1947 and 1948 as a part of a stud)- of the behavior and ecology of the Arctic Terns •Steraa paradisaeayof MachiasSeal Island, Bay of Fundy, New Brunswick. Canada. Although a greater portion of time had to be allotted to behaviorstudies, sufficient data have been acquired to allowan analysis of eggand chickmortality of the ArcticTern and to providematerial for discussionsof the factorsaffecting group success in thisbreeding colony. Thisis fortunatesince Dr. RaymondA. Payntermade detailed studiesof the eggand chickmortality of HerringGu,11s (La. rus argea- tatussmithsonianus) and or.herspecies on Kent Island,only fifteen milesaway, during the summersof 1947 and 1948 t Paynter1949). Thus,by coincidence,so•ne comparison of data obtainedin the same locality,during the sameyears and for a relatedspecies, is possible. Due to the exceptionallywide distributionof the Arctic Tern, in- formationon the speciesis scatteredthroughout the ornithological ,literatureof manycountries and of manyyears. In this paper,an attempthas been made to bringtogether much of the moreimportant informationdealing with the ecologyof the species. I wishto expressmy appreciat.ionto Dr. RalphS. Palmerfor many helpfulsuggestions based on his experiencewith terns. Drs. RobertT. Clausen,Richard H. Backusand CharlesL. Remingtonhave assisted withidentification of plant, fish and invertebrate specimens. My wife assistedconstantly with field work which could hardly have been completedwithout her. SPRING ARRIVAL For manyyears, the lightkeeperand his wife, Mr. and Mrs. Ottawa Benson,kept recordsof the arrival datesof the Arctic Terns at Machia• SealIsland. Arrivaldates and comments from their log for fivecon- secutiveseasons are presentedbelow.

57 Bird-Bandi•ng 58] H•WKSLEY,Breeding Population o.fArctic Terns April

In 1945,the terns arrived on May 17 and startedlaying on May 29. In 1946,they also arrived on May 17. In 1947,the terns arrived on May12 but did not alight on the island until May 20. Ternswere first seenon May 12 againin 1948but did not alight until May 17. In 1949, thebirds were heard on May 11 andfirst seen on May 12. Theylanded brieflyon May 14, but the presence of a DuckHawk (Falco peregrinus) anatum)kept them off for a coupleof days.After the hawk was killed by the keeper,the ternsreturned and eggs were found by May 29. Althoughthe terns arrived on May 12, 1948,there were only a few eggslaid whenI arrivedon June 3. We receivedthe impression that the numberof birds on t,he islandincreased for severaldays after that, but sawno arrivingflocks. Accordingto the Bensons,the ternsarrive at MacbiasSeal Island in large,high-flying flocks. The first time they appear,they circle high above the island, then leave. Afterthat they fly overthe island each morning in socialflight but do not alightuntil theyhave done this for severaldays. Pahner11941:401 has described thislast type of behaviorsimilarly for the CommonTern (Ster•mh. hirundo). THE ENVIRONMENT FOOD AND FOOD-GETTING The rich waterssurrounding Macbias Seal. Island provide an abundant foodsupply for thatternery. ArcticTerns, like otherspecies of terns, probablycolonize areas close to goodfood supplies.This doesnot necessarilymean that the ternery must be locatedclose to the sea. Colonies are known to exist even in the interior of the Scandinavian countriest Kullenberg 1946: 12t wherethe birds fish in lakesand rivers. In Bail:inLand, Dalgety (1936• notedthat Arctic Terns "appearedto preferfishing in freshwater." In contrastingterns with gull,s,Tinbergen (1932: 11-12) remarks itrans. 1: In manyways the gullsappear remarkably adaptable, much more sothan the terns. The gullslook for food in many ways,changing with the circumstances.The terns, on the other hand, show a strar•gepsychic unadaptability. Their behavior,especially on their searchfor food, is very stiff. We could call them, briefly, over- specializedplungers. Althoughthis conceptis fundamentallycorrect, one mustnot think of the Arctic Tern as a wholly unadaptablespecies. It is principallya deep-waterfeeder, yet it quicklyadapts to the useof shall.owowaterfood organismsor of insectsat timeswhen these are especiallyabundant. In a colonyfive milesfrom the sea,Wright (1909) observedthat the youngwere being fed not with fish,but with largecrane-flies and may- flies. These terns were caught for banding on hook and line baited with crane-fliesand "greendrake." Palmer (1941: 17) did not see pelletsfrom Commonor RoseateTerns (Sterna dougalli), but I fre- quentl.ynoticed pellets, containing insect parts, on the rocksin the roost- ing areas on MachiasSeal Island. The food habits of the Arctic Tern have been little studied. A com- pilationof specificfood organismswhich the speciesis knownto utilize the world over would be of litt.le value in this study and might even Vol.1957XXVIII H•WKSLEY,Breeding Population o[- Arctic Terns [59

give an incorrectimpression. Specific names will be usedonly in referringto food usedin the studyarea. A surveyof the literatureindicates that the three mostimportant food groupsinclude fish, crustaceansand insects. Fish from the families Gasterosteidae,Ammodytidae, Stromateidae, Scorpaenidae, Cyclopteridae,Blenniidae, and Gadidaeare much used. Crustacean groupsmuch used include branchiopods, copepods, mysids, amphipods, euphausids,and decapods.Insect orders represented are Ephemerop- tera. Odonata,Lepidoptera, Coleoptera, Hymenoptera and Diptera. A few pelagicmolluscs such as Clio, Limaci,a and .Loligoare eaten, as well as marine worms. Availabil.ity is obviouslya very importantfactor in whatfood is used. In Labrador, O. L. Austin, Jr. (1932: 128• did not seeArctic Terns eat anythingbut the lanceAmmodytes americana, but thesesha•llow water fishwere not usedat all at MachiasSeal Island. CaptainFielden (1877) madethe interestingobservation in polar regions,where continual ice preventsfish from comingto the surface,that the amphipodAnonyx nugaxworks its way up throughtidal ice cracksand is eatenin numbers by the Arctic Tern. The fishesmost used for food by the terns at Machias Seal Island werehake (Urophysissp., probably U. chuss),the 1.umpfish(Cyclopterus lumpus) and the dollar fish (Poronotustriacanthus). However, the supplyof skeletonshrimps (Meganyctiphanes norvegica) and of lump- fish appearedto be mostconstant. Thesewere used extensively during intervalswhen hake and dollar fish seemedto be lessplentiful and were usedto someextent throughout the entire season. There werenoticeable seasonal "runs" of mostof the food organisms. Notableamong these were the "runs" of cicadas(Oka,agana cana•4en- sisy an annelid (Nereis pelagica), small squids (Loligo pealei) and dollar fish. The cicadaswere used during the secondweek of July in both 1947 and 1948. In 1947, Nereiswas used from July 8 to July 11, squidscame into usefrom July 10 to July 17, and dollar fish were used in large numbersfrom July 24 until we left the islandin early August. Other fish, in the order of the frequencywith which they were brought to the colony, were four-'beardedrockling (Enchelyopus cimbrius),Clupea sp., haddock(Melanogrammus aegehfinus), rock eel (Pholis gunnellus) , three-spinedstickleback (Gasterosteus aculeatus) , rosefish (Sebastosmarinus), and barrelfish (Palinurichthysperei- ]ormis). The nine-spinedstickleback (Pu,gitius pungitius) was commonly usedat coloniesnear Churchill,Manitoba, during studiesmade there in 1949. Unidentifiablemoths and a dragonfly(Tetragoneuria spinigera) were found in nestswith young.

EFFECTS OF WEATHER AND LIGHT The importanceof weatherto ArcticTerns is indicatedin manyparts of this paper. Weatherconditions have a direct effectupon the ability of adults to seek food. Later it will be shown to what extent weather affectsthe survival of the young. Roostingor brooding adults may becomeso wet in rain or densefog that they are unableto fly out of tal,1 60] [la;;•sLEY,Breeding Poptdation o.iArctic Terns Bird-E•andingApril

lOO

• 70

• 60 • so

2 3 4 5 5 20 25 3031 1 July Aug.

Fig. 1. Growth of chicks as related to •eather.

vegetationand may be easilycaught. Weatherand light influencethe amountof socialflying to so•nedegree. Weather, especiallythat which affects visibility, and the need for brooding young, has a very decided effect upon the ability of Arctic Ternsto providefood for their young. Weightsof chicksfrom July 2 to August1, 1948, showeda definitecorrelation between certain weather conditionsand the averageweight of the group of chicks. Figure 1 is a graph showingthe averagedaily gains and lossesin weight and the visibility factors which seemedto contrt•l them. It is interestingto note that a singleday of fog or unfavorableweather may not resultin weight losses,but that several days of unfavorable weather cause a definite weight loss. It becomesunderstandable that a continuationof adverseweather of this sort might greatly decreasethe survival of young. Kullenberg (1946) is probably correct in thinking that the absenceof the Arctic Tern as a breedingbird on the easternAsiatic coast south of the Tchuktch Peninsu•la and at the Sea of Okhotsk is to someextent due to the great frequencyof fog there. The Arctic Tern is generallycredited with seeingmore daylight duringa year than any otherspecies, but detailedinvestigations of its daily rhythmas correlatedwith light in the Arctic regions,and histo- logicalstudies to searchfor specialadaptations in the retinaof the eye, havestill to beattelnpted. Pahngren (1935) found that birds in Lapland 'Vol.1957xx'vIII H•'W•SL,•:'•, Hreeding Population o./•lrctic Terns [61 were mostinactive between 6 and 11 p.m. This period of rest followed the hottestand dryestpart of the day. Severalauthors have mentioned feedingthroughout the night by Arctic Terns in high latitudesand at MachiasSeal. Island the ternsfly aroundto someextent even during the hoursof darknesswhen they cannotsee to feed. A.J. Marshall (1938• mentionsperiods of quiescencefor SO,hespecies, including the Arctic Tern, even during the perpetuallight of the Arctic, but alsomentions that theseare affectedby weatherconditions. The observerwho spendsan extendedperiod of time in a ternery cmmothelp but be awareof the spectaculartype of socialflight which is oftenperformed by ternsduring the sunri• or sunsethours. Thistype of flightis closelyallied to the morningand evening flights performedover the ternerybefore it is occupied.It may be considered as a temporaryreversion to flockingbehavior which probablyserves part.lyto maintainthe socialorganization of the colonyso necessary for communaldefense. It is not entirelyclear, however,what servesas the releaserfor thesemass flights. In 1947,I receivedthe impressionthat socialflights reached a peakof intensityat timesof specialstress and excitement.Some of the mostintense social flying occurredjust on the"eve" of thehatching of mostof the youngin thecolony, and again at thetime when many of the younghad begun to fly. However,weather and light intensityseem also to be involvedand in order t.ocheck this possibility,I kept daily recordsof eveningsocial flights and weather in 1948from June21 to August2. FromJune 21 until.July 13 the amountof eveningsocial flying recorded each evening checked closely withthe weatherfor that evening.Social flights were large on clear evenings,less intense on partly overcastevenings and did not occuror werehardly noticeable on eveningswhen there was rain or fog. The onl•exception occurred on the evening of July2 whenno socialflights wereseen although the sky was clear. However,at that time,the adults wereespecially busy feeding newly hatched chicks right up untildark- nessdescended and this drive may havebeen stronger than that of flocking. FromJuly 14 untilAugust 2 therewas a frequentlack of evening socialflying even in goodweather, but it shou,ldbe keptin mindthat theyoung required increasingly greater amounts of foodduring this periodand that frequentperiods of bad weathermade fishing more difficult for the adults. Lightappears to alsohave a definiteconnection with morning social flights.and with the incubationdrive. The followingincident accom- panieda changein the amountof light at 10:15a.m., July 1, 1948,on MachiasSeal Island. The skygrew dark and fog camein on a sudden breeze. Within a few lnomentsthe light readingon a Westonmeter droppedfrom 400 to 25. All eggsand chicksin the areawere immedi- atelycovered by adults.Five minutes later the fog had passed and the skybecame bright again. With this,a seriesof outflightsoccurred muchas theywould have during social. flight at sunrise. Eveningsocial flights occur at varioustimes and show little close correlationto thetime of sunset.If the skysuddenly darkens in late afternoon,then brightens again, the flights may occur early. On c.lear Bird-Banding 62] H4WKNLEY,Breeding Population of .d,'ctic Terns April dayswith normal sunsets, flights occur about the timethat the sun appearsbrightest on thehorizon. Morning social flights always occur closeto the time of sunriseand are lessoften prevented by dull skies thanare eveningsocial flights. This I believeto be dueto the extreme changefrom darkness to daylightwhich occurs in spiteof weather.It is quitepossible that not only a suddenincrease of light,but suddenly increasedlight precededby relativedarkness, stimulates social flying. Theseobesrvations agree rather well with Armstrong'sstate•nents (1954) that the general.activity levels of birds may be affectedby changesin light-intensityand elevation of thesun and that bird activity "seemsto be moreclosely geared to the light-intensitiesin the morning than in the evening."Brackbill (1952) has madeobservations which suggesta correlationbetween departure times of waterfowl and light intensities. GROUND AND VEGETATION The Arctic Tern prefersnesting sites with lessvegetation than do other northernnesting tern species.Gravel bars, sandybeaches and outer islandswith little or low vegetationare typical breedingplaces. The use of outersea islands by the Arctic Tern contrastssharply with themore frequent use by theCommon Tern of islandsin mouthsof rivers and in bays. Suomalainen• 1939 t believesthat the intensereflection of light.from the seaon gravelbeaches where it nestsis importantto the Arctic Tern. However,O. L. Austin, Jr. t 1929t pointedout what is perhapsthe most important factor in the relationshipbetween terns and the vegetation in their nestingcolonies, that is, the wing and tarsuslengths of the speciesin qu,estion.He foundthat RoseateTerns, which have relatively short wings and long tarsi, are well adaptedto comparativelydense growthsof grassand shrubbery. CommonTerns have mediumlength wings and tarsi and are fairly well adaptedto either open spacesor rather denselycovered areas. The Arctic Tern, with its long wingsand short tarsi, is least well adaptedto high vegetation,but prefers to nest in openplaces where it can fly to its nest or at leastto the immediate vicinity of its nest. An Arctic Tern colony at Hopkins Island, Massa- chusetts,ceased to flourish when the island became overgrown with bushes fAustin, Sr. 1940). In view of these relationships,it is interestingto note that Arctic Terns rather than CommonTerns occupythe most denselyvegetated part of the nestingarea on MacbiasSeal Island. This situationoccurs becausethe j u•mbledrocky area most usedby CommonTerns on the island is irregular enoughto be undesirablefor use by Arctic Terns, which accordinglyhave preemptedthe open turf locationsmore de- sirable for them. However, as the seasonadvances, these open turf areasproduce some of the tallestvegetation on the iMand. If Arctic Ternsare surprisedwhile on the groundin high vegetation, they often may be caughtby hand. Nest sitesare chosenso earl}- in the seasonthat there is no indicationof high vegetationat the time the site is selected. Individuals which becomethus hamperedby vegeta• tion hold their wingsabove them in readinessfor flight more than do birds nestingon open ground. [• grassand yarrow .• kis • Rumexcrispus, Ro mexicanus •-• grassand Rumex acetosella • grasses •yarrow, Umbellfferae,docks • Umbelliferae,docks • scatteredyarrow, Sedum [---• rocks • fence

ß walk

Limit oœsoil area

High tide line

Fig. 2. Distributionof coverplants on MachiasSeal Island,New Bru,n,swick, duri,ng summers of •947 and 194,8. 64] HAWKSLEY,Breeding Population o/Arctic Terns Bird-BandingApril

MachiasSeal Island is granitic with a thin layer of peaty topsoilon the central portion where most of the vegetationgrows. An attempt was made to collectall the speciesof vascularplants occurring on the island. No woodyplants occur on the 15-acreisland. In two years, 41 speciesof plants from 18 families were collectedand preserved. Fourteenspecies were Europeanintroductions. These European forms probabl.y cameto the island in hay broughtthere for domesticanimals. Only one of these,Rumex acetosella,furnishes extensive ground cover in tern nestingareas. The yarrow (,4chillea lanulosa) is the most im- portantpl. ant on the islandfroin the standpointof controllinghabitat. Rume• acetosellais next in importanceand Iris versicolorcontrols a reasonablylarge area. Figure 2 indicatesthe distributionof the important coverplants and groupsof plants on the island. Terns nested on nearly all the vegetation-coveredareas except those closest to build- ings, and nest areas extendedout a short distanceonto rocky areas with sparsevegetation. Near-byGull Rock has only two vascularplants. They are ,4grostis stoloni]era i dwarfed) and Plantago maritima. Both grow in cracks in the rocks. The isletsoccupied by breedingArctic Terns at Churchill,Manitoba, were usuallylocated in lakes or ponds. The most commonplant com- munity on theseislets was a Cladonium-Ericaceaecommunity in which Ledurn decumbenswas very abundant (Shelford and Twomey, 1941). Onesuch island which has supported an Arctic Tern popu,lationof about 20-30pairs for manyyears measured only 16 by 50 yardsand lay but 24 yardsfroin the nearestpoint on the shoreof the lake.

WATER The Arctic Tern is probablyless dependent.upon water for food than is the CommonTern. Even thoughthe SugarloafIslands (Maine) "wereteeming with grasshoppers"in July, 1938, Palmer (1941: 24) did not see the CommonTerns there catch any. He believesthat Common Ternsusua_l, ly take insectsonly from the surfaceof the water and not from the air or on land. The Arctic Tern seemsto be more adaptable. There are numerousaccounts of Arctic Terns hawking insects. One such accountis given by Roberts (1934: 256) who observedArctic Terns in Iceland taking numbersof mothsin this manner. The moths were capturedat heightsof about 10 to 20 feet in the air. One stomach contained56 mothsand as many as 60 per cent of the terns present were feedingon moths. Robertsalso reports a spider(Lycosa tarsalis) froin onestomach and smallamounts of grass(Agrostis palustris) from two stolnachs.He believedthat the birdshad pickedup the grasswhile huntingfor insectson the ground. On numerousoccasions I saw both adults and pre-flying juvenals drink froin fresh-waterpool. s on Machias Seal Island. These same poolswere usedfor bathingby a few juvenalsbut the adultsregularly resortedto .saltwater to bathe. Adult Arctic Terns at Churchill, Mani- toba, bathein fresh-waterponds near the terneries. Both fresh and salt waterappear to be suitablefor this purpose. Vol.1957XXVIII HAWKSLEY,Breeding Population o/ Arctic Terns [65

Duringmigration at sea,or duringpelagic wanderings, the birds must either drink sea water or obtain sufficient water from their food. Studieson this subjecthave not beenmade. ArcticTerns alight on floatingobjects at seain preferenceto alighting on the water but are buoyantand swim well whenin the water. The lengthof time they can remain on the water without becomingwater- loggedhas not been determined.

ASSOCIATES A varietyof otherspecies has been reported as nestingin closeasso- ciation with Arctic Terns. These associates include birds from other familiesas well as numerousspecies of Laridae. In Alaska,the Glaucous-wingedGull (Larus glaucescens),the Short- billed Gull (Larus canusbrachyrhynchus), the PacificKittiwake (Rissa tridactylapollcarls), Sabine's Gull (Xema Sabini) and the Aleutian Tern (Sterna aleutica), comprisethe Laridae most often found nesting in companywith Arctic Terns. An Alaskancolony of 2'3pairs of Arctic Ternswas unmolested by a pair of Short-billedGull.s nesting near them according to R. B. Williams i1947). Bailey (19271 reports that Arctic Terns in Alaska con- tinually harrassedthe Short-.billedGulls. In a colony of 60 to 75 pairsof AleutianTerns with a few pairs of Arctic Ternsamong them, Friedmann• 1933) thoughtthat the AleutianTerns seemedmore selec- tive in preferringto nestby themselves.Howell (1948), on the other hand,reports no segregationin a mixedcolony of AleutianTerns and Arctic Terns in which Pacific Kittiwakes also nested. On the eastern coast of North America, the Common Tern, Roseate Tern, andLeast Tern (Sternaalbi/rons), nestwith the Arctic Tern as far north as their breedingranges extend. In 1948, two pairs of Laughing Gulls (Larus atricilla) nestedamong the terns on MachiasSeal Island (O. and J. Hawksley1949). Thesegu•11s were unmercifullyharrassed by the terns each time they attemptedto alight near their nests,but onceon the groundthey werenot disturbed.One pair raisedtwo young. the other pair deserted. A singlepair of Herring Gulls nestedon Gull Rock, adjacent to Machias Seal Island, in 1948 and 1949, but neither Herring Gulls nor Great Black-backedgulls (Larus marinus), were toleratedon MacbiasSeal Island proper except below the tide line on the north end of the island. This was the part of the islandcl. osest to Gull Rock, where the gulls regularly roostedand sunned. In 1948, about 20 pairs of Arctic Ternsattempted to neston Gull Rockbut failed. This mayhave been due to maraudingby the gullsbut no evidenceof such destruction was found. Ross'sGull (Rhodostethiarosea), nestswith Arctic Terns in Siberia t Buturlin19061. The firstnest of thisgull discoveredin Greenlandwas in the midstof a colonyof Arctic Terns (Dalgleish1886). The Marpies • 1934) studiedtern coloniesin whichnot only Common Terns and Roseate Terns nested with Arctic Terns but also Sandwich Terns (Thalasseuss. sandvicensis),and Least Terns. 66] HAWKSLEY,Breeding Population o/ Arctic Terns Bird-Banding^prn

A most interestingecological rel.ationship exists between Sandwich Ternsand the gullswith whichthey associate.This has beenreported on by Salomonsen(1947). The terns nestin the centerof the colony and are surroundedby Black-headedGulls (Larus ridibundus ridibundus). CommonGulls (Larus canuscanus), are in turn scatteredat the edgeof the Black-headedGulls, and Herring Gulls (L. a. argentatus), are permittedonly amongthe more scatteredne•ts of the CommonGull. The terns nest late and force themselves into the center of the Black- headedGull concentrationby destroyingeggs and drivingout gulls. Nevertheless,the Bl.ack-headedGulls are aggressiveand may givesome protectionto other speciesin the colony. This brings fort.h the questionof commensalismin nestingcolonies of Laridae and relationshipsof Arctic Terns to speciesoutside the Laridae. Smallbirds suchas sandpipersare usuallyignored by terns. Spotted Sandpipers(Actitis macularia) often wanderedamong the nests of terns on Machias Seal Island. They were chasedby the terns only at the time whentern eggswere closeto hatchingand incubatingbirds wereexceptionally nervous. Important associatesof the terns on Machias Sea] Island by sheer weightof numbers,were approximately400 pairs of CommonPuffins (Fraterculaau arctica)•and 2,000 pairs of Leach'sPetrels (Oceanodroma leucorhoa leucorhoa). However, there seemsto be no direct relation- ship betweenthese species and the terns in that colony. Williamson (1948) reportedthat Arctic Terns in the FaeroeIslands gleaned the fishdropped by puffins,but that they did notmolest the incoming puffins andseemed to beeating most of thefood on the spot rather than carrying it to their young. I never saw terns scavengingin this manneron Machias Seal. Island. Red Phalaropes(Phalaropus [ulicarius) occupiedan island (about 100 x 50 yards in size) in Spitzbergenon which Arctic Terns also nested{,Padget-Wilkes 1922). Sutton (1932: 192) reportedMandt's Guillemots (Cepphus grylle mandti), Common Eiders (Somateria mollisimaborealis), and Old-squaws(Clangula hyemalis), nesting with Arctic Ternson a smallisland near SouthamptonIsland. The statusof speciesof Anatidaewhich nest in gull andtern colonies hasbeen much discussed. The TuftedDuck (Aythya ]uligulai is most notedfor thishabit in northernEurope and the Old-squaw is commonly found nestingwith Arctic Terns in North America. Fabriciu.s(1938} andHaartman (1938) bothfelt that Tuf•ted Ducks preferred to nestwith Laridaebecause of theprotection thus afforded them from predation of HoodedCrows (Corvus c. cornix)upon their eggs. The most important factpointing to commensalismis the opensituation of the nestsof these ducksin gulland tern coloniesas opposedto concealedsituations when theynest away from gulls and terns. I haveobserved this same phe- nomenonwith Old-squaws nesting in the Churchill,Manitoba, region. Onthe basis of observationsin Finland, Bergman (1941) believesthese associationsare basedon social,instinct rather than on protection. Durango'simportant paper (1949), however,does not bear this out. Vol.1957XXVIII HaWKbLEY,Breeding Population o/- Arctic Terns [67

He mentionsthat Tufted Ducksand Ruddy Turnstones(Arenaria i. interpresjchange their breeding grounds to followchanges made by CommonTerns •/nd Arctic Terns. He feelsthat the colonyof terns is a releaserfor the Turnstoneand thinks that both hereditary and learning elementsguide the birdsto their choiceof a nestingsite. He further pointsout that in England,where they breed rat.her late and on ponds andlakes with densevegetation, Tufted Ducks do notnest with Laridae. Thusthe probabilityof puresociability seems doubtful and it is more likely that the Tufted Duckswhich neston isletswith scantvegetation in Swedenand Finland,do sobecause they are lessdisturbed by enemies there. There is a hint of mutualismin Sutton's (1932: 192) statement that he found at least six Arctic Tern nestson one island "which appearedto havebeen made in old nest.sof the Old-.squaw." The occurrenceof Arctic Tern eggs in the nestsof other birds is interestingbut probablyof little significancein most cases. Sugden i1947) found exotic eggs in nestsof California Gulls (Larus cali- /ornicus1and felt that the gul,ls carried them t,o the nest to eat, but were overcomeby broodingimpulses instead. The gulls supposedly carry the eggsin their bills. Bailey •,1925) found a Gl,aucousGull's (Larus hyperboreus)nest which containedone egg of its own, one of an Arctic Tern and one of an Old-squaw. The gull may have acquired thesetwo eggsin the way Sugdensuggested. The Arctic Tern egg which Bailey •1926) found in a BlackobelliedPlover's (Squatarola squatarola)nest was undoubtedlylaid there by a tern. It is common to find clutchescontaining eggs of both CommonTerns and Arctic Ternsin mixedcolonies. On MachiasSeal Island, an Arctic Tern egg somehowcame to rest in a burrow with an egg of Leach'sPetrel (J. Hawksley 1950). Mammalianassociates which cannot. be classedas predatorsare not numerous.In the outer Hebrides,Gray Seals (Halichoerusgrypus) frequentsotne of the rocksused by Arctic Ternsbut pr.obablydo not enterthe nestingarea. At Churchill,I foundthat ,boththe Muskeg MeadowMouse (Microtus pennsylvanicus drummondii) and the Varying Letoming (Dicrostonyx groenlandicusrichardsoni) lived on the islets in tundra pondswhere Arctic Terns nested,but saw no evidenceto indicatethat either harmedeggs or youngterns.

PREDATORS Relativelyfew animalsare able to catchand kill adult Arctic Terns. Thegreat majority of specieswhich may be consideredpredators of the ArcticTern prey uponthe eggsand nestlingsor uponjuvenal birds whichare lessadept at escapingon the wing. The line of distinction betweenparasitic and predatory species is ratherfine, especially where suchbirds as jaegersare involved. Jaegersmight be classedas social parasites,but they are definitelypredators at times. Exceptat the southernlimit of its range,snakes are not important predatorsof the Arctic Tern. Over-abundantsnakes (Thamnophis ordinatus)were credited by Floyd (1932) with consumingeggs and Bird-Banding 68] HAWKSLEY,Breeding Population o[ Arctic Terns April youngof ternson PenikeseIsland, but snakesare seldomfound on the outer islands where Arctic Terns nest farther north. Avian predatorsare probablythe most important. These consist chieflyof threetypes: jaegers, hawks, and otherLaridae. AkhoughElfrig (1905) mentionsfinding bones and feathersin the stomachsof Parasitic Jaegers (Stercorarius parasiticus) these were probablyof smallbirds for it is doubtfulthat jaegerscould seriously injure ternsor gulls. The weaknessof the ParasiticJaeger's bill and feet has beenpointed out by McCabeand Racey (1944). They show that its ability to terrify its victim with psychologicalwarfare which "smothersthe flight" of the victim is the importantfeature. The killing of an Arctic Tern by ParasiticJaegers is reportedby Roberts•1934), but deathwas the result of the crashingtogether of four jaegersas they convergedon a singletern which had a fish. Jaegersare more effectiveagainst terns when several gang up on one tern, but two or three terns may easilydrive off a jaeger. It was felt by Trevor-Battye(1897) that neither a Parasitic Jaegernor a Long- tailedJaeger (Stercorarius longicauda) had a chancewhen opposed by a pair of Arctic Terns. Jaegersalso foll. ow migratingArctic Terns. Long-tailedJaegers were notedtraveling in companywith them by Wil.lett (1918) in Alaska. Falconsare amongthe few birds which are actually able to catch and kill adult Arctic Terns. Duck Hawks were the most important Fredat.ors on adult terns at MachiasSeal. Island. Thesefalcons pre- sumablycame from the islandof GrandManan and made the trip rather frequently. In spiteof the massdefense which was made by the whole colony,the Duck Hawksdid succeedin taking terns. I found remains of an adulttern, whichhad undoubtedlybeen eaten by a falcon,on Gull Rockin 1948. Mr. Benson,the lighthousekeeper, usually shot at Duck Hawks which came near the island and once shot a tern from the talons of one. Duck Hawkswere blamed by Norton (1907) for the abandon- mentof Libby Island,Maine, by terns,but damageby DuckHawks at MachiasSeal Island seemsrelatively slight and the falconsshould not be molested. At Churchill,I sawa PigeonHawk (Falcocolumbarius columbarius) mobbedby ArcticTerns. Rawcliffe(1949) reportsthe takingof a youngArctic Tern by a Kestrel (Falco tinnunculustinnunculus). Dr. Arthur A. Allen informs me that Gyrfalcons(Falco rusticolus)in Alaska take Arctic Terns and Sabine's Gul!,s. TheLaridae which prey upon Arctic Terns include most of thelarge speciesof gullswhich breed in the samerange. Sabine'sGull. like the jaegers,tends to bemore a parasitethan a predator.It is sodependent upon the Arctic Tern in Greenlandthat it fails to nest when conditions areunsuitable for theterns (Bird and Bird 1940). It is alsofrequently seenmigrating in companywith Arctic Terns and jaegers. Akhoughlarger species of gul.lssuch as the Glaucous.Iceland, Glaucous-winged,GreatBlack-backed andHerring Gulls are generally concededto prey upon terns to someextent, little quantitative work has beenattempted todetermine theextent of their depredations. Thegen- Vol.1957XXVIII HAWKSLEY,Breeding Population o!Arctic Terns [69 eral effectof larger gullsupon the dynamicsof a tern populationwill be discussed later. Owls and crows may sometimesprey upon Arctic Terns but are probablynot seriouspredators since there is little to indicateso in the literature. Eastern Cr•ows(Corvus b. brachyrhynchos)were mobbed and driven off by the terns at Machias Seal Island. Kirkman t1908• cites Howard Saunders on an incident in which Arctic Terns mobbed and drowned a Hooded Crow. Unusualand seriousavian predatorson MachiasSeal Island were a few semi-domesticatedMallards (Anas platyrhynchos).Mr. Bensonfirst discoveredthat the duckswere eatingeggs and newlyhatched young of Arctic Terns. A few days later, while in a blind, I saw two ducks in action. They werego,bbling nearly-hatched eggs as quicklyas they could find them in spiteof viciousattacks by the terns. They musthave eaten at leasta dozenbefore I couldget out of the blind to chasethem. After that, the duckswere fed more liberally in an attemptto preventfurther damageto the terns. Althoughthe slaughterof terns for the feathertrade no longercon• tinues, man is still a serious predator upon terns. Fortunately, the birds are small enoughnot to be very attractiveas food under most conditions. However,tern eggsare usuall.yvery wel•lliked by humans living within the rangeof the Arctic Tern. Where peopleimpose some regulationof eggingupon themselves,as they often do in the Bay of Fundy,the actualharm donemay be smallor no harm at all may result. Underprimitive conditions, people such as Eskimosand Indiansare less inclinedto regulatetheir egging. They do not objectto eatingeggs in advancedstages of incubationand thustake the eggswhenever they can find them. Such a situationis cited by Salomonsen(1955) for the Gr•0nneIslands in Greenlandwhere at least100,000 Arctic Tern eggs were collectedeach season,causing a decreasein the species. O.L. Austin,Jr. (19321 citesa casein whichone Eskimo girl took800 eggs at one time from a ternery in Labrador. In his extensivepaper on the subjectof egging,Cott t 1953-1954)lists the ArcticTern as beingamong the groupof birdssecond most heavily exploitedfor eggs. He further liststhe Arctic Tern as "showingde- cline,or egged-outlocally," but feelsthat the effectof eggingon bird populationsin the Arctic areas of the Old World is not clear. The peopleof the GrandManan region consider tern eggsto be a delicacyand coll'ectthem whenever possible. More importantto them, of course,is thecollecting of gull,eggs. Thiswas a ratherwell-regulated processin the past. Ownersof islandsoften rented them by the day to partiesof eggers.The natives claim that Herring Gulls lay upto nine eggsif eggsare taken as soonas they are l'aid. They claim to have always left the l•astfew for the birds. Egginghas been conducted on MachiasSeal Island for yearsin spite of thefact that it is a sanctuary.This has been done mainly by the keepersor their guests.The local"rule" is that eggsmay be collected until clutchesof two are found. This simplerul•e is enforcedby the conditionof the eggsafter that, so that eggingonly lastsfor abouta weekin the first part of June. Bird-Banding 70] I-[AWKSLEY,Breeding Population o/ ,4rctic Terns April I doubtthat any harm co•nes from this early egging. If anything,it mighttend to tnakethe time of hatchingcome a bit later and thus youngchicks would be lesslikely to die in stormswhich seem to be frequentin lateJune and early July. At least,the MachiasSeal Island colonyhas flourishedin spiteof egging. Cott (1953-1954)lists a number of birds which are probably similarly benefitedby egging. A•nongthem is the Lapwing(Vanellus vanellus) in Holland. Tern eggswere used for food duringthe first Worl'dWar in Great Britain and researchhas beencarried on in Britain recentlyto determine the palatabilityof the eggsof birdswith a view to possiblefuture con- sumption.Egg tasters at the ResearchStation, Cambridge, have given eggsof the Arctic Tern a rating of 6.5 which is in the "relatively palatable"group. A rating of 10.0 is ideal and 2.0 is inedible. The DronesticFowl has a top rating of 8.6, the Com•nonTern rates 7.3 and the Herring Gull 7.9, all threebeing in the "highly palatable"group I Cott 19491. Someof the da•nagedone by humansin terneries,such as thosealong the coastsof Maine and Canada,is not true predation, but a form of vandalism. Picnickersand sight-seerswho visit terneriesinadvertently stepon eggsand chicks. They pick up and tnisplacechicks which are later killed by adultsfor trespassingon adjoining territories. If they bring dogsto a ternery,untold darnage may result. The •1ogkept by the fatnily at MachiasSeal' Island wastrained to stay away from the tern nestingareas, but former keepershad dogsnot so well trained. Sheepand cowskept in terneriesstep on tnany eggsand young. A cow on Machias Seat Island picked up and badly mauled juvenal terns. There were no wild mammal.s living on Machias Seal Island except bats (Myotis). Farther south,however, Norway Rats (Rattusnorvegi- cus) are seriouspredators in tern col.onies. O. L. Austin, Sr. (1948a) has reportedextensively on this predatorin the CapeCod colonies. A report by Forbush (1921) statedthat a stnall colony of Arctic Terns at Chatham,Massachusetts, was wipedout by high tidesand later by raids of "catsand skunks,which destroyed both eggsand youngand drovethe parentbirds away." In the Arctic, foxestnay be seriouspredators upon terns. In 1949, Mrs. T. C. Stanwel•l-Fletcher and Miss Hazel Ellis found no less than five densof the Arctic Fox (,/llopexlagopus) on the Fox Islandsnear Churchill. Therewere perhaps a thousandpairs of ternsnesting there and it appearedthat eggs,at least,were being taken by the foxes. The fox pupswere extretnely curious and foil.owed the humansabout. When attackedby theterns, the pups jmnped into the air in attetnptsto strike back. In NortheastGreenland, Ermines (Mustela ermiaea) preventedthe breedingof ArcticTerns on Ternholme Island in 1938according to Bird and Bird (1941). The Arctic Tern'shabit of frequentlynesting singly on the tnainland in arcticregions tnay be effectivein reducingpredation, especially from matronallanpredators such as the Arctic Fox. Vol.1957XXVIII HAWKSLEY,Breeding Population o/ Arctic Terns [71

INVERTEBRATE PARASITES Relativelyfew invertebrateparasites are knownfrom the Arctic Tern. In searchingfor Mallophagaon adultsand chicksI wasnot alwaysable to find any on the birds. Only a few individualson MachiasSeal Island were found with heavy infestationsof lice. No macroscopicendopara- sites were noticed. Slater and Carter (1886) found youngArctic Terns in Iceland dying in the downystage soon after hatchingdue to either "the cold summer or to the presenceof huge tapeworms"with which they were infested. A heterophyidtrematode (Cryptocotyle lingua) is knownto infectgulls, ternsand otherpiscivores •Willey and Stunkard1942 i. Anothertrema- rode (Aporchis rugosus)is definitely a parasite of the Arctic Tern • Linton 19281. The most interestinghost-parasite relationships are those between the ternsand their Mallophaga.During the summerof 1947, I collected all the MallophagaI found on Arctic Terns. They were eventuall'ysent to Englandfor identificat.ion. Theselice provedto be of two new species,Quadraceps houri Hopkins, (Hopkins,1949• and Saemundssonia lockleyiClay •Clay,1949a). Fourteenof theMachias Seal Island speci- mensbecame paratypes of the new Quadraceps. The CommonTern hasthe sametwo generaof lice, but two different species.The); are Quadracepssellatus and Saemuadssoniasternae. This is interestingin view of the fact that all three speciesof terns of the genusChlidonias have the samespecies of Mallophaga,Saemunds- sonialobaticeps •Clay 1949b). Even more importantis the fact that Saemundsso•ialockleyi is found not onl.yon the Arctic 'Fern but also on the Antarctic Tern (Sterna vittata). Miss Clay •1948t believesthat this suggestsa closerelationship between the two hostsand supports Kullenberg'ssupposition •1946: 78• that the Antarctic Tern is a relatively recentderivation from populationsof the Arctic Tern which re- mainedin the antarcticand sub-antarctic;,:aters during the northern summer. Gross(1937} reportsa speciesof Philopterustaken from an Arctic Tern. There are a few other referencesin ornithologicalliterature to Mal,lophagafrom Arctic Terns,but the identificationswould bear check- ing now that two new specieshave beendescribed for that host. One mite was taken from an Arctic Tern at Machias Seal Island and proved to be Euhaemogamasusoudemansi. EGG AND CHICK MORTALITY A knowledgeof mortalityprior to fledgingis essentialto an under- standingof the reproductivecapacity and managementof a speciesor of populationproblems in general. MacbiasSeal Isl.andis especially well suitedfor a study of natural mortality becauseit is relativelyfree from disturbingoutside influences. The environmenthas been only slightlyaltered by the presenceof manand no unnatural,predators have beenintroduced except a few domesticanimals which are fairly well controlled. The term "fledging"has been variously used and mightlead to con- fusionif not defined. In this paper,the term is usedto refer to the age of attainingflight. 72] H•,•KSLEY,Breeding Population o.[Arctic Terns Bird-Bandin,gApril

Large-scalebanding of youngwas carried out in the summersof 1947, 1948, and 1949. A continuationof this bandingand additional trappingof adultsin the summerof 1950was done by GlenWoolfenden. The methodof studywas essentially the samein both1947 and 1948. Four areaswere used, two in eachyear. Owing to the pressureof other work, thoseused in 1948 containedfewer nestsbut were carefully chosen so as to be representativeof typical breedingareas on the island. One area in 1947 included 100 nests which were marked with num- bered woodenstakes on June 14. The 100 nestsrepresented all the nestscontaining one or more eggswhich couldbe found in that area on the day of marking. This assureda random assortmentof clutch sizes. The area was located at the northern end of the island and in- cludedpart of a beachof smoothpebbl.es. The remainder,and greater part of the area. was on peaty ground made irregular by tussocksof vegetationand by burrowsof Leach'sPetrel. The predominantvege- tation was a yarrow (Achillea lanulosa). The second area used in 1947 contained 50 nests marked on June 23. This area was locatedon higher ground, nearer to the center of the island. It was coveredby short grassesand Rumexacetosella except on its outer i seawardl edge where there was someyarrow. It includeda few somewhatl'ater nests than the firstarea because again •I1 nestswith eggsin the area were used. In • 948, one area with 50 nests and one with 25 nests were marked. The 50-nestarea, which was marked on June 7, was located on the westside .of the islandbetween the dwellingsand the sea. The inner sectionof the areahad somegrasses, but mostof the area wasrocky withonly small amounts of yarrowand Sedum roseurn growing where there were patchesof soil. The 25-nestarea was locatednearer the southernend of theisland, west of the "whistlehouse." The presence of a blind in this area made more detailed observationsof the occu- pantsof the areapossibl. e. Anotherdifference in this studyarea was thefact that nests were marked as theywere established. The laying processand markingbegan on June4 and wasnot completeduntil June13. All nestswere in adjacentterritories. Thenests in eacharea were checked daily. Whenthe young hatched, theywere banded with color bands until oldenough to holdregular aluminum bird bands. In addition, notes were taken on the col'orsof the soft parts and on the down of chicks. When chicksbecame older andbegan to wander,nests were checked in bothmorning and evening. In thisway few chicks were lost to thestudy. It is reasonab,leto assume thatthose which did disappear died, because the island was thoroughly coveredduring banding operations and had any of thesechicks sur- vivedthey would probably have been identified immediately.

CLUTCH-SIZE Lack (1947} givesthe usualclutch size for the Arctic Tern as two for both Englandand Norway. The numbervaries from one to three. In a laterpart of thesame paper, Lack (1948) givesthe average clutch as 2• eggsin mid-Europe.Bickerton (1909) studiedthe clutch size of the ArcticTern in theBritish Isles and reported 355 eggsin 209 Vol1957 XX¾III H4WK.EY,Breeding Population o]Arctic Terns [73

nests.This givesan averageof 1.7 eggsper nest. The differencebe- tweenthese figures and thoseof Lack may well be connectedwith the fact that Lack's are "basedon generalstatements" and that someof his authoritieswere oi51ogists.I have noteda tendencyamong oi51og• iststo assumethat the normalclutch for this speciesis two and to avoid collectingthe one eggclutches which they suspectof being incomplete. Data on clutch size tend to be more accurate if taken from studies suchas lPettingill'st19391, and my own studyon egg and chick mortality, for then the historyof the clutchis known. Pettingill's100 nests on Machias Seal Island contained144 eggs,or an averageof approximately 1.4 per nest. In 225 nestsstudied on the sameisland, I found 308 eggs,or an averageagain of approximately1.4. Eklund 119441 checked279 nests in Greenlandwhich averaged 1.7 eggsper nestand R. B. Williams •1947• reportsan Al.askan colony of 45 nestsin which all of the clutcheswere of 2 eggs. It would appear that the averageclutch, at leastin North America,is smallerthan indi- catedby Lack. Clutchesof four, five, or moreeggs have been reported by many authors. Call 11891'i discussedsome sets of four to six but notedthat mostshowed clearly that they had beenlaid by two birds. The 50-nestplot of 1947, whichincluded later nests,showed no sta- tisticallysignificant difference in clutchsize when comparedto the earlier 100-nestplot. The averageclutch size for the 50-nestplot was 1.16 -- .05 and that for the 100-nestplot was 1.20 •- .04. The 50-nest plot of 1948 had an averageclutch size of 1.76 • .07 comparedto 1.68 • .11 for the 25-nestpl. ot. This again is an insignificantdiffer- ence,showing that the averageclutch size was approximatelythe same regardlessof the fact that one group was picked at random after one or more eggswere laid and the other groupwas picked as the first egg was laid in each nest. In a study of Herring Gull and LesserBlack-backed Gull (Larus ]uscusgraellsij colonies,Darling (1938: 62• found that the average clutchesin 1936 were 2.1 for Herring Gullsand 2.2 for the LesserBlack- backedGulls. In the sa•necolonies the following year, the number of birds present doubled and the average clutch size was 2.9 for both species.Darling concludedthat thesefigures pointed to "an improve- ment in the environmentalcomplex in 1937." The differenceon MachiasSeal Island, betweenan averagedutch sizeof 1.19 in 1947 and of 1.73 in 1948 mayhave a simil.ar significance but clearcutreasons are not evident. There werethree knownchanges in the Machias Seal Island colony in 1948 from the conditionsin 1947. In 1948 the weatherwas generally less severe, the heightof the vegetation was lower and the breedingpopulation had increasedto some extent. However,the collectiveeffect of thesechanges, ear,ly in the seasonwhen eggs were laid, couldnot havebeen great. Darl.ing worked with relativelysmall populationsin which changesin numberswould be very noticeable,whereas the populationincrease at MachiasSeal Islandwas relatively slight compared to the total population. Thetheory, that some birds in thesouthern part of theirrange lay fewer eggsthan those in thenorth and that the increased length of dayin the northallows the parentsto collectmore food to feedmore young, has 74] HAWKSLEY,Breeding Population of Arctic Terns Igird-Bandir•gApril

lOO

ß ß ß

80 ß

;

.../

\ \ \ \

\ \

30 Averageofsingle chicks - Average of fraternal chicks

Average of all chicks 20

•o .... • .... io is 2'o 25 Age in Days

Fig. 3. Chick growth as related to brood-size. Vol.1957XXVIII I•AWK•LEY,Breeding Population ofArctic Terns [75 beenrevived by Lack (1947). Lack did not presentany data indicating that this was true for the Laridae. Paynter •1949) statesthat "it is probablethat it doesnot existin this groupsince . . . the survivalrate of the youngseems to be in.dependent of the numberof chicksin the nest and greater food-gatheringtime would not increasethe chances /or survival." AlthoughPaynter's data for the Kent Island Herring Gulls support his statement,my own data from the main study (see Tables V and VIi and from a supplementarystudy of MachiasSeal Island Arctic Terns,do not agreewith this generalstatement for the Laridae. Un- fortunatelythe number of nestswith which I was able to work in the supplementarystudy was s•nall, but I believe that the data show a definitetrend which might be even more clear with more data, and thereforeinclude this study. In 1948, while taking routineweights of chicks,I was able to check to seewhat differencethere was betweenthe developmentof single chicksand chickswhich competed for food with anotherchick in the samenest. Accordingly,the routineweighing of chickswas expanded into a supplementarystudy to provide comparativesurvival data. Weight gainsand lossesof chickswere correlatedwith weathercondi- tionsas alreadyshown in figure 1. Twelvenests were selected for thisstudy but onlyeight of thempro- ducedchicks. Five of the nestsproduced two chickseach and three producedsingle chicks. Figure 3 showsthe results. Weights of single chicks,weights of fraternal.chicks, and the average weight for all chicks areplotted for comparisonfrom hatching to an ageof 22 days. The fraternalchicks' weights represent only those still competing at thesame nest on any given day. Fromthe fourteenth day on, only onenest still hadtwo chicksalive. The youngerchick in that nestdied at an ageof 22 days,at whichtime it weighedonly 46.2 grams.The otherchick in thenest was then down to 63.6grams but eventually recovered and lived. The averageweight curve for fraternal chickswhich survivedwould comeclose to the averagefor all chicks,and is thereforenot shownin the graph,to preventconfusion. Sincethe numberof nestsstudied is small,these results are not con- clusivebut they indicate that the survival rate of theyoung may very wellbe affectedby the numberof chicksin the nest,at leastin someof theLaridae. No stu,diesof thesurvival of youngArc,tic Terns appear to havebeen made in any latitudehigher than that of MachiasSeal Island,but thereis someindication that the clutchsize may average larger. TableI summarizesthe accurateaverage clutch sizes obtainable for pointsenough farther north than MachiasSeal Islandto have con- siderablymore summersunlight. However, these colonies are repre- sentedonly by datafrom singleyears or evensingle days. Also,as is pointedout in Paynter's(1954: 139) paperon Tree Swallows(Irido- procuebicolor), the really significantdifference in length of day is basedmore on the time of nestingthan on latitude. Birdsnesting further north do so nearer to the summer solstice than those farther south. 76] HAWKSLEY,Breeding Population o./Arctic Terns Bird-BandingApril

TABLEI. 2lyerage clutches /or •4rctic Tern at latitudes higher than Macbias Seal Island.

Locality Nests Clutch Source British Isles 209 1.7 Bickerton (1909) Bear Is., Barents Sea 49 2.0 Duffey and Sergeant •1950) Greenland, lat. 62 ø 279 1.7 Eklund (1944) Southampton Is. 127 1.8 Sutton (1932) S. E. Alaska 45 2.0 R.B. Williams i1947)

DURATION OF INCUBATION In a paperon a "biologicaldistinction" between Common and Arctic Terns, Van Oordt (19341 stated that the Arctic Tern does not begin incubationuntil the clutchis comp,letewhile the CommonTern begins whenthe first egg is laid. I have found no suchclear-•ut distinction. Arctic Ternshatch at intervalsof a day or morejust as CommonTerns do. Only occasionallydo two chicksin the samenest hatch on the same day. Table II showsthe historiesof laying and hatchingat nine nests. All nestsin the 25-neststudy area whichhad two eggs,and in which botheggs hatched, are included.Birds in thisarea were observed dail.y from a blind so that the actions of most of the adults were known. The figurefor the secondegg in e•tchnest represents incubation time for that egg. As indicated,the first eggmay or may not havebeen incu- batedfor the numberof daysit wasin the nest.

TABLEII. Intervalsbetween hatching and laying o/ individualmarked eggs o/ the •4rctic Tern.

Interval in Days, Nest Days in Nest Laying Hatching Remarks on No. Egg 1 Egg 2 1 & 2 1 & 2 Incubation 3 24 21 4 1 Began with 2 6 22 22 1 1 .... 1 8 23 20 4 1 .... 2 13 23 21 2 0 .... 2 15 23 22 2 1 .... 2 16 21 22 1 2 .... 1' 18 24 23 2 1 .... 2 19 22 22 2 2 .... 1 X 26 26 2 2 .... 1•' * Chick wandered. ? Late nester; birds very nervous.

It may be readily seen,even in this small sample,that not all birds delay incubationuntil' the clutchis complete.Those which do, tend to coverthe nestat irregularinterval. s until the clutchis complete.This accountsfor thefact t. hat the firstegg usually gets about one day ahead of the secondin thesecases. In the table,egg 2 in nest16 wasincubated longerthan egg 1 becausethe wanderingof the firstchick had kept the broodingbird off the nestfor muchof the last two days. Vol.1957XXVIII [-tAWKSLEY,Breeding Population o/ Arctic Terns [77

NestX representeda late nesting.The "nervousness"of the birds resultedin rather intermittentincubation. Excludingthis nest,the average period of incubat.ionfor secondeggs in the other8 nestswas about 22 days. Pleske •1928: 224) arrived at this same figure by calcula- tions basedon the time that Co•nmon'Tern eggs required in an incubator (Evans1891). The period is givenby Witherby et al (19411 as 21-22 days. Besidethe lengtheningof the incubationperiod due to the irregu- 1.arityand "nervousness"of the birds, somevariations may be caused by thenesting habitat itself. SomeArctic Terns nest in barerocky areas. The rocks retain the sun's heat and act as a natural incubator when the eggsare not covered. Other individualsnest in shadedareas where the sun's heat is much less.

HATCHING SUCCESS Of the 308 eggsin the four main studies,111 failed to hatch. This givesa hatchingsuccess of slightlyless than 64 per cent. Nestswhich had complete or partiMhatching success are analyzed in TableIII. Thosenests which contained only one egg had a hatching successof 59.3 per cent while those containing t.wo eggs had a hatching successof 68.9 per cent. Appl.icationof the Chi-squaretest for the significanceof these frequencies shows that they might occur by chance aloneand therefore cannot be consideredsignificant. The figures for nestswith t,hree eggs represent too few datato be significantbut are includedin the tablefor completeness.

TABLEIII. Clutch-sizerelated to /ailure to hatch.

Hatched No. Eggs Total No. in Nest of Eggs No. P•r Cent 1 144 85 59.3 2 158 109 68.9 3 6 3 50.0

Totals 308 197 64.0

A summaryof causesfor the failure of eggsto hatchis given in TableIV. Thereis little in the literatureon the fertihtyof ArcticTern eggs.Montague •1926) examinedbetween forty and fifty clutches."Of the clutcheswhich contained two eggs,just underfifty per centcon- tainedone infertil'e egg. Of the singlecl. utches, a smallproportion were infertile." Pettingilli1939) found5.5 per centof 144 eggs"sterile." Sincethe eggs are not examined until the end of theincubation period in mortalitystudies such as Pettingill's,it is nearly impossibleto dis- tinguishinfertile eggs from eggs which become addled early in develop- ment.Accordingl'y, I have designated eggs ini:ubated for a longperiod but failingto hatchas "infertileor addledearly'" unless an identifiable embryowas present. Embryo deaths include those chicks which died in hatching.Punctured eggs appeared tobe punctured bythe terns them- 78] HAWKSLEY,Breeding Population o/Arctic Terns Bird-BandingApril Vol.1957XXVIII H4WKSLEY,Breeding Population o/ Arctic Terns [79 selves.Some punctures were of a sizethat couldhave been made by the birds' bill,s, but somewere small and may have been made by the toes of the birds when they flew off suddenly. Most of the eggsindi- catedas having been "destroyedby man or domesticanimal's" were steppedon by four sheepwhich frequentedthe 50-nestplot in 1948.

FLEDGING SUCCESS Of the 308 eggsin the four plots, 197, or 64 per cent, hatched. A daily search was made for each chick and its history was recorded until it died or was fledged. Seldomwas a chick which was missing for over five daysfound alive. Young chicksmissing for that long a period could rather safel.ybe assumedto be dead. The averageage at fledging varied in the four areas from about 24 days to about 30 days. Robust chicks which were well leath- ered and ready for flight were consideredfledged when they dis- appearedat ages cl.ose to 30 days. Many of the young birds were actual.lyobserved in their first flightsand manywere caught after they had flown. The first primariesof a few birds,caught just after a first flight,were measured to determinethe length(flattened). The primaries of birds nearly able to fly were then measuredand the measure- mentsrecorded with their histories.This providedan additionalcheck on whetheror not a bird of fledgingage had actuallyflown when it disappeared.Birds having primaries over 100 min. longwere considered ableto fly if light enoughand strongenough. AllChicks

First Chicks

SecondChicks

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Age in Days at Death

Fig. 4. Brood-size as related to survival. Bird-Bandir•g 80] HAWKSLEY,Breeding Population o.[Arctic Terns April

A histogram(figure 4-A) showsthe agesat whichchicks died. It canbe seenthat over76 per centof the deathsoccurred during the first weekof life. The heavymortality shown for the third and fourth daysrepresents deaths of secondchicks (see figure 4-D), in nestswith two chicks. First chicksin nestsof two, and singlechicks also show highmortality during the firstweek (figure 4-B and C} but not asheavy as that of secondchicks. The highermortality of secondchicks is due to increasedcompetition as the older chicksin the nestsbecame dis- proportionatelylarger and strongerdue to their head start. Another factor is that the presenceof two youngchicks in a nest often means that the parentsmust both spenda greatdeal of their time fishingand consequently,two chicknests are seldomguarded by an adult (as single chicknests are} while the secondadult is off fishing. Since a test has been made to determi•e whether clutch-size affects hatching success,it will a.lsobe of interest to test for the effects of clutch-sizeand brood-sizeon post-hatchingmortality. This may be doneby coinparingthe number of youngfledged from one-egg nests and one-chicknests with the numberfledged from two-eggand two-chick nests.The basicdata are shownin TablesV and VI. Fledgingof one chickper nestis consideredto constitutesuccessful fledging.

TABLEV. Clutch-size as related to fledging success.

No. Eggs Total No. No. Young YoungFledged in Nest Eggs Fledged Per Nest 1 144 54 .38 2 158 52 .66

All nests 308 107 .48

A Chi-squareanalysis of thesedata showsthat the frequenciesare significant.There is lessthan one chancein 100 that the variationin

TABLE¾I. Brood-size as related to fledging success.

No. Young Total No. No. Young Percentof in Nest Young Fledged YoungFledged ! 106 75 70.8 2 88 31 35.2

All nests 197 107 54.3 fledgingsuccess could have arisen by chance. It may be concluded that althoughclutch size has no significant effect on hatching success, both clutch-sizeand brood-sizeshow an ultimateeffect in the successof fledging.One-chick nests have better hatching success, but two-egg nestsgive better fledging results than one-egg nests. This seeming in- congruityis partlyexpl'ained by somereduction of two-eggclutches to onebefore hatching. Parental age, as related to clutch-sizeand to successfulbreeding, may give otherexplanations when such a factoris studied in Arctic Terns. vol.1957XXVHI HaWK•L•:•,Breeding Popttlation o/ ,4rctic Terns [8].

These results differ notably from those obtained with the Herring Gull t Paynter, 1949} in which three-eggnests had the best hatching successand in which "surival of the young, at least until the thirtieth day, is independentof the brood-size." This indicatesthat it would be difficult to make general.statements on the significanceof clutch-size in the Laridaefrom data on one or two speciesonly. It may alsopoint to a basicdifference between gulls and terns,between a speciesmuch subjectto early predationand one little subjectto suchpredation, or to a differencebetween species laying clutchesaveraging less than two and thosespecies .laying larger clutcheson a whole. This cannot be determineduntil morespecies within the family havebeen investigated. The causesof chick death are summarizedin Table IV, but. need some elaboration here. A large proportionof the 33 deathsby starvationoccurred in two- chicknests. The youngerchick was usually the victimdue to its in- ability to compete. Twenty-threechicks disappeared. There was little doubt that these died. Pettingill(1939} had29 chicksdisappear in his studyof Arctic Ternmortality on MachiasSeal Island. Thisis a ratherhigh figure for only 100 nests,and raisesa questionas to the amountof disturbance whichoccurred in the studyarea. He attributedthe lossof thesechicks to "kidnappingor wanderlust." One chickwas steppedon by a cow, and four were steppedon by sheep. Onecow was on the islandin 1947,and four sheepreplaced it in 1948. Seven died due to exposurein periods of cold rain. Those which drownedfell into poolssurrounded by steeprock sides which prevented their climbingout. The one chick designatedas "premature"still had a large yolk sac when completelyhatched, even though it hatchedon the twenty-second day of incubation. In all, 107 chicksfledged. This givesa successof only 34.7 per cent if figured on the total number of eggsbut represents47.6 per cent successin terms of pairs which succeededin fledginga chick. The latteris a morereasonable type of figurefor a specieswhich seenis to be capableof raisingonly one chickper breedingpair per season. The figureof 34.7per cent may .be used, however, to comparewith Pettingill's very low figure 11939) of 15.9 per cent successfrom 144 eggsin 100 nests. Pettingillgives three reasonsfor this low figure (27.3 per cent was the lowestfigure for any one area in my studies): 1) maraudingby adult terns, 2• failure of nestingdrive, especiallyin young breeders, and 3• coinpetitionfor food. The first reasonis logical enoughand well supported. In evaluatingthe second,however, it must be pointed out that Peningill'sstudy did not begin until Jul.y 2, when he "selected ß . . one hundrednests with eggsstill unhatched." At this late date in the season,many of the nestingbirds which still had eggsmay not only have included young adults in which the nesting drive was "incomplete,"but would also have includedsecond nesting attempts in whichthe birds had passedthe peak of the nestingdrive. Thus, the 82] HAWKbLEY,Breeding Population o.[Arctic Terns Bird-BandingApril

nestsselected were hardly typical of the entire seasonin that colony. Further,the statedspan of his stayon the islandis lessthan incubation plusthe periodfor attainingflight in the ArcticTern. Competitionof adults (as implied) for food seemsunlikely because the surrounding watersare rich and the terns adapt themsel.vesquickly to changesof supply. Whenone food supply fails, the birdsturn to oneor severalof the otherabundant food sources.O. L. Austin,Sr. (1946) statesthat all writers,including himself, "have failed to appendcredible, sub- stantiatingdata" when reporting food shortages affecting chick mortality in tern colonies. The summerof 1948was comparatively free of stormsand fog, but thesummer of 1947produced a bad rainstorm with cold northerly winds on July 4 and 5. In additionto this, the summerof 1947 was extremely foggy. Oneperiod of fogin Julylasted for 400 hours.The generally betterweather conditions of the 1948season may account partially for the betternesting success of that.year. Paynter•1949) foundthat 36.8 per cent of theeggs in hisstudy of KentIsland Herring Gull. s produced fledged young. In thesame year, 1947,I foundthat 37.1 per cent of theeggs in thestudy of MachiasSeal IslandArctic Terns produced fledged young. Similarityof weather conditionsmay well have been invo.lved in the likeness of these figures andthe generalsuccess of fledgingbecause weather undoubtedly has indirecteffects as well as the direct effect of killing young.

POPULATION VARIATION AND MANAGEMENT

NON-BREEDING AND DESERTION Many authors,including O. L. Austin, Sr. f1934), have suggested that tern populationsare cyclic in numbers,but little informationhas beenpresented by anyonebut Austin on this matter. One of the most interestingphenomena bearing on this subjectis that of non-breeding. Many caseshave been reportedin which Arctic Terns were presentat arctic breedinggrounds during the breeding seasonbut were preventedfrom nestingor made no attemptto nest. The followingare examples. In 1855 Evans and Sturge •18591 found the Arctic Tern to be common in WesternSpitzbergen but its eggswere not found that year. Manniche 11910• reportedthat terns and someother speciesdid not nestin northeastGreenland in 1907. Bird and Bird 11935) reported the speciesas present on Jan Mayen Island in 1934, but the birds "madeno attemptto breed." The Birds i19411 mentionnon-breeding of the Arctic Tern in northeast Greenland in 1938 but believed that predationby Ermines preventedthe speciesfrom breeding. Late arrival,and as a consequence,the possibilitythat the birdshad passed the peak of breedingcondition was probablythe causeof the very reducedbreeding reported by Seligmanand Willcox •1940• for Jan MayenIsland in 1938. J. G. Williamsi1941• reportedthat all the breedingcolonies on the Varanger Peninsula,East Finmark, were deserted in the late, cold summer cf 1939. Vol195i XX¾III HqWK,•LEX,Breeding Population o/ Arctic Terns [83

A considerationof the factors involved led Lack (1933) to the con- clusionthat periodicnon-breeding of the typeexhibited by the Arctic Tern is dueto "late seasons"in whichsuitable nesting sites are covered with snowand ice until the gonadshave regressed to the point where thebirds are no longercapable of breedingthat. season. This conclusion is borne out by more recentobservations such as thoseof the Birds •1940• and A. J. Marshall (1952). This non-breedingcannot be con- sideredthen, as a truly cyclicphenomenon. Desertion has been known to occur in tern colonies due to excessive egg robbing or due to the continuedpresence of certain predators. Where terns nest in large colonies,they have littl.e successif they are hamperedby numerouspredators and oftendesert. Another type of desertion,however, has been describedby N. Marshall 11942} who observed it in Common Tern colonies on Lake Erie. The desertiontook placein the form of extendedevening social flightsin which nearly the entire populationtook part and from which they did not return to their nestsuntil the followingmorning. Marshall wascompletely puzzled by this performancewhich he sawtake place severaltimes in the years 1939, 1940 and 1941. The colonyhe studied at StarveIsland was unsuccessful.all three seasons.In 1939, the 1,052 nestsof May 25 producedonly a singlechick. The populationdropped offto 513pairs in 1940and to 109pairs in 1941with only a fewyoung raisedeach year. Chillingof the eggsduring the periodsof night desertionmay have beenresponsible for the lack of success. Marshallwas apparent,ly unaware that eveningsocial flights in which the birds return, are normal, al.mostdaily, occurrencesin terneries. If, as he seemed reasonably sure, neither human disturbance nor predatorswere causes, only one explanation seems plausible in the light of our presentknowledge of tern behavior. That is, that for some reason,the intensityof the socialflight increased[o sucha point that social behavior dominated individual behavior so that the terns went off in a flock insteadof returningto incubation. If the breederswere predominantlyyoung, their incubatingdrive may have been more readily overcomeby a flockingdrive (seeO. L. Austin, Sr., 1945). On June 27, 1948, I witnesseda peculiarsocial flight of Arctic Terns at Machias Seal Island which had some of the characteristics of the desertionflights described by Marshall,but the birds finally returned to their nests. Out-flightsbegan on the north end of the island at 12:50 p.m. Thesesoon spread to mostof the islandand took on the form of the morning socialflight. The day was clear and bright and upon searchingthe sky for a predatorwhich could have causedt.he alarm we saw nothing. At 1:05 p.m., a group of about 100 terns alightedin a raft on the water about one-half.mile off-shore. They did not bathe, but a few momentslater arose en masse and flew back to the island. It wasnot until 1:30 p.m. that the terneryquieted down to continueits normal activities. Sincethis phenomenalflight occurred just prior to the time of the h,atching of mostof the young,it may be that it wasa resultof pent-uptension in the incubatingbirds. I have no definiteevidence for this idea,however, and it wouldnot explain similarlypuzzling flights which occurred at othertimes. Theseflights Bird-l•andin,g 84] HAWKSLEY,Breeding Population o.[Arctic Terns April are muchlike thosewhich the Marpies I1934: 169) called"panics." Althoughin neitherthis case nor in Marshall'sobservation, any predator was seen,this doesnot excludethe possiblepresence of one,since out- flightsare typicallyset off by visualor auditoryalarm stimuli.

ECOLOGICAL SUCCESSION AND OTHER HABITAT CHANGES Plantsuccession and othervegetational changes have a definiteeffect on tern populations,either directly, or indirectlythrough consequent changesin populationsof otherspecies of birds inhabitingthe area. Many environmentalchanges have occurredduring the past 50 years in areasthat havebeen used by breedingterns. Most of thesechanges have been due to changesin land use rather than to true succession. Sincethey have often beenabrupt and drasticchanges, their effecton tern populationshas been more disastrousthan the normal ecological successionwould have been, for, in many cases,the outer islands on which Arctic Terns nestedhad rather stabilized ecologicalconditions until man interfered. In 1932,thousands of youngor eggsof CommonTerns and Roseate Terns were destroyedor mysteriouslydisappeared from Penikese Island,Massachusetts. Floyd (1932t thoughtthat the disappearanceof eelgrass(Zostera marinai from the surroundingwaters at that time may have causeda failure of food for the terns, but there was no con- clusive evidence for this. Vegetationalchanges on islandsin the CapeCod region have had a very clear connectionwith popul'ationchanges. I have already cited the instancei Austin, Sr. 1940) in which the overgrowthof Hopkins Island by bushesforced out a flourishing colony of Arctic Terns. Reclamation i by humans) for terns of some of these areas has been accomplishedby destroyingthe high vegetation. The Marpies (1934) refer to the encroachmentupon tern nesting territoryby Black-headedGulls and the resultingmoves made by the terns. The terns either move to an entirely new area, or if there is sufficientroom, they move to anotherpart of the samearea, apart from the gulls. The increasein numbersof Herring Gull•sand Great Black-backed Gullsand the decreaseor shiftingof tern populationsalong our eastern seaboardhas beena matter of concernto many studentsof birds in recent years. In statingthat the CommonTern "musthave an inherentability to makereadjustments to the necessitiesof alteredenvironment," O. L. Austin, Sr. (1942) must have referred mainly to situationsin which theterns were not in directcompetition with gulls. Crowelland Crowell •1946t have shownthat on rel'ativelysmall islands, such as those in the Weepecketgroup •Mass.l, whereterns must competewith gulls, the terns are forced out as the gulls increase. The successionof nestingspecies on MuskegetIsland, Massachusetts, has furnishedan outstandingexample of populationchange. Early in the nineteenthcentury, Muskegetwas primarily a tern colony and probablythe largestsuch colony on the New Englandcoast.. Later, the LaughingGull appearedthere and by 1850 becamean abundant Vol.1957XXVIII H a.WKSLEY, Breeding Popttlation of Arctic Terns breeder. Its numbersfluctuated, due to eggingand the feathertrade, but in 1945, 16,000-22,000pairs were estimatedto be present. Since that time, Herring Gulls,which began nesting on the islandin 1922, havebecome so numerous that therewas a reductionof LaughingGulls to about2,500 pairs in 1947. The ternsthen nesting on Muskegethum; bered only about 1,000 and were limited to one area (Gross19481. The LaughingGull is oftenfound nesting with terns,but usuallythis associationoccurs on islandswith high growing herbs and thus they are islandsmore used by CommonTerns and RoseateTerns than by Arctic Terns. The LaughingGull is more tern-like in its movements than the larger gulls and due to its manoeuverabilityit is more readily able to colonizein terneries. Once it becomesestablished, however, the way is openedfor intrusionby the larger speciesof gul,ls. The nestingof two pairs of LaughingGul.ls on MachiasSeal Island in 1948 might not have occurredif the island had not had so much high vegetation. Some of the changeswhich have taken place on MachiasSeal Island have a directbearing on populationsthere and will now be discussed. The estimatesof Arctic Tern popul'ationson MachiasSeal Island have largely been guesseswhich appear to have been influencedby the guessesof former visitorsto the island. Estimatesof about2,000 pairs havebeen given frotn 1911 •Brown) until 1937 (Pett.ingil.1 1939) but it seemsunlikely that the population would have been that constant. It is apparent from details given to me by the keeper that it has varied greatly in recent years. When Mr. Bens.onbecmne keeper in 1944, there wereabout 40 sheepon the island. Thesebelonged to 'the former keeperand sincethey couldnot be taken off until Augustor September,the terns failed to raise any young that year. Needless to say,the vegetationon the islandwas close-cropped and muchof it was destroyedcompletely. In 1945, the birds returned and nestedsuccess- fully and by 1946 they were presentin larger numbersthan Benson had everseen them, although he had workedas assistantkeeper or had visitedthe islandduring a peri.odof manyyears. In 1946, the vegetation was still low, but by 1947 it reacheda height of about40 inches, whichcaused the permanent residents to talk aboutthe islandbeing the "weediest"they had seenit. This was probably due to the lack of grazing animals l exceptfor one cow), and to the fertilizer contributed by the 40 sheepwhich were taken off in 1944. In 1948, a declinein the island vegetationbegan l with the aid of foursheep l andthe yarrowwas four to six incliesshorter at blooming timethan in 1947. Foursheep were still presentin 1949and the yarrow wasonly a few incheshigh whenwe arrived on July 6. It then had onlyabout two moreweeks to reachmaximum height which probably was muchlower than in 1948. In addition,much yarrow had been suppl'antedby Rumexacetosella and grasses. Severalmethods of makinga censusof the breedingArctic Terns on MachiasSeal Island in 1947and 1948were considered but only one methodseemed practical. This wasan indirectmethod which used data obtainedfrom the mortality studies and "saturation" banding of young. Themajority of birdsbanded were banded close to fledgingage when 86] HAV•K•LEY,Brcediag Popttlatioa o/ ,4rctic Terns lgird-lgandi•gApril mortalitywas slight. Bandingwas done systematicallyby movingup and down marked lines which extended from the center of the island to the edge of the nesting area. When a completecircuit of the island had been made in this manner, a second circuit was made and so on. With each succeedingcircuit of the isl'and,fewer young werebanded until the bandingsper unit of effort approachedalmost zero and it was estimatedthat practicallyall young on the island had beenbanded. During the processof banding,bands were removed from thefe• deadchicks found and placed on live chicks. The total breeding adultpopulation was then figured by usingthe number of chicksfledged by the total pairsnesting in the survival.study areas as an index. In 1947, the number of juvenalsbanded was 1,281 which were all thatcould be found. Fromthis andthe factthat 150 pairsin the study areasfledged 66 young,it wascalculated that. the breeding adult population wasapproximately 2,900 pairs. A total of 1,592 juvenalswere bandedin 1948 but we could have bandedabout 300 more if we had had more bands. Since 75 pairs fledged41 young,this indicatedthat the total breedingadult population was around 3,450 pairs. The limitations of statisticalreliability of such figures are obvious but they do allow somemeans, better than guessing,for comparingthe populationsof the two years. Our party banded1,384 juvenalsin early July, 1949, and Dr. and Mrs. SouthgateHoyt banded308 more in the latter part of the month. Since our three-daybanding effort by no means includedall birds bandableat that time, and sincesome birds fledgedbefore the Hoyts arrived,I am reasonablysure that the colonywas as large or laTgerin 1949 than it was in 1948. The weather in 1949 was much like that of 1948and success in rearingyoung seemed to be good. Duringthe fali and winterfollowing a bandingseason, AHison Ben- son,son of the keeper,searched for bandson skeletonsor on crippled young left on the island. He found 22 bands after the 1947 season and 27 after the 1948 season. Most of these bands were taken f•'om juvenalswhich had brokenwings. They had probablyreceived these injuriesin flying,so this wouldnot alterthe populationestimates. The steadyincrease in the sizeof the breedingpopulation from 1945 through1949 mighthave been prevented if the high vegetationof 1947 had continuedto persistor if the LaughingGulls which nestedon the islandin 1948 had renestedand increasedthe foil.owingyear. The vegetativnalchange from 1947to 1948showed up in the greatincrease of nestingmaterial in nestsof 1948. Thematerial was mostly stems of deadyarrow left from the year before. It wasalso noticed that some 1948nests were formed of ringsof driedbits of sheepdung. CommonTerns showed some increase during the th'reeyears. In 1947 a total of 42 youngwere banded. CommonTerns were not banded afterthe band supply ran low in 1948,but 40 hadbeen banded up until that time. A total of 110 young CommonTerns were bandedon the islandin 1949. TheCommon Tern has pressed back and replaced the ArcticTern in several1.ocalities in GreatBritain (Robinson,1920, vol.]9•?xx v•l• IIAWK•LE•,Breeding Polmiatio• of . •4rcticTerns [8•

1921a,1921b) but it is unlikelythat this will occuron MacbiasSeal Islandunless the vegetationincreases for a periodof severalyears.

"INTERNAL FACTORS" Thusfar, mainly"external factors," or factorsof the outsideenviron- ment, have been consideredas affectingtern populations. "Internal factors," or factors within the behavior of the speciesitself, are also important. The certain thresholdof numbersnecessary for successfulbreeding in many speciesof Laridae [Darling 1938• is a factor of this sort. Palmer •1941: 101) has already mentionedthat Arctic Terns which nestsingly on t•e tundra near Churchill,Manitoba, gather together for socialflights which probablyresult in necessarysensory stimulation. Darling'sidea that the larger the colony,the earlier and moresuccess- ful the breeding,was borne out by the attemptedbreeding of Arctic Terns on Gull Rock in 1948. Thesebirds nestedlater •sti.11laying on June 20• than the birds on Machias Seal Island and did not reach numbersexceeding 20 pairs. By July 16 all eggshad disappearedfrom Gull Rockand no youngwere found. Thesebirds were cl.ose enough to take part in social flights with birds from the main island so their failure may have been due partly to small numbersand partly' to the presenceof a pair of nestingHerring Gulls and otherroosting gulls on the smaller island. Townsend(1923) tells of a populationof 1,000 pairs of terns on this rock, equal to the number breeding on the main island at that time. These were either driven off by gulls or otherwisefound conditions better on the main is[and for no terns have nested on Gull Rock for sometime. Pettingill [1936) found that six pairs of Herring Gulls [presumablyin 1935) had replacedthe ternson this rock and thought that they would have a bad effect on terns on the main island if not checked. The inabilityof ternsto competewitk gullsis thought.by Tinbergen (1932) to be dueto the unadaptabilityof ternsunder changing conditions. Thereare other"internal factors" which affect population, such as infertility, breakageof eggsand maltreatmentof chicks,but these are of small importanceindividually as comparedwith the basic psychicmake-up of the birdsas mentioned by Tinbergen. Oneother "internal factor" which is importantis theage composition of the breeding population. Austin, Sr. •1945) has shown for the Common Tern that ,the behavior pattern and successof the whole colony dependsto some degreeupon the influenceof older breeders whichhave nested in the samelocality over a periodof time.

MANAGEMENT Theprobability of themanagement of sea birds to obtaintheir eggs as foodfor humanconsumption is not entirelyremote, but mostman- agementthat takespl, ace in the near futuremore likely will be for aestheticreasons. Management of this typehas already been carried outwith success bythe Austin Ornithological Research Station on Cape Bird-Ban(ii•;g 88] HAWK•LE•,Breeding Population o/ .drctic Terns April

Cod. Tern breedingareas which had becomeovergrown were at first reclaimedby re•novingthe vegetationby hand, althoughplowing, harrowingand raking were recognized as a betteranswer. Later, tractors wereemployed in this work,much of whichtook placeon Tern Island which is sandy. The terns respondedby decreasingnest concentration, thus decreasing•nortality due to internal strife and allowing•nore area for nestingin general. In addition, debris, wreckageand shackswhich harboredrats were eliminatedand red squill was plantedgenerously. (Austin, Sr. 1934, 1940, 1946; Floyd 1938). Hundredsof terns, including10 pairs of Arctic Terns,nested on a new area createdby dredgin.g at CapeCod in 1947 (Austin Sr., 1948b). Managementof the MachiasSeal Island ternery is unnecessaryexcept for the possiblefencing-in of dronesticanimals which show signs of doingdamage to the terns. The centralpart of the island l seefigure 2) has •nuchgrass which couldbe usedby grazinganimals. This area is not usedby terns,sheep or cowsbecause of the presenceof humansin the central area, but enclosedanimals would be forced to use it. Managementat Matinicus Rock, Maine, would be beneficial. U.S. CoastGuard personnel allowed cats to run free thereduring the second World War. These cats became feral and were credited with the destructionof the largetern colonywhich formerly existed there. It is saidthat• rats also inhabit the island. Exterminationof thesepredators shouldeventually bring back the terns,but to my knowledgesuch a program has not been undertaken.

SUMMARY Arctic Ternsare principallydeep water feedersbut adaptreadily to changesin food supply. Insectsbecome important in the diet when abundant. Weatherand light affectfood procuringand socialactivity. The Arctic Tern dependsless on waterfor food than the CmnmonTern, but mostof its food is takenfrom the sea. Water is alsoimportant for bathing. Nestingsites with littlevegetation are mostdesirable. A changeto highvegetation in a breedingarea will oftenprevent Arctic Terns from breedingin that spot,either directlyfor mechanicalreasons. or in- directlyby favoringspecies better adapted to highervegetation. Associatesinclude a greatvariety of othercharadriiform and anseri- form birds. Thereis evidenceof commensalismin which Old-squaws and Tufted Ducks,especially, are benefitedby nestingin Arctic Tern colonies. Importantpredators include gulls, hawks, jaegers and some•nammals. Jaegersand Sabine'sGull tend toward being parasiticon the Arctic Tern. The Mallophagaof the Arctic Tern indicatethat it is •noreclosely related to the Antarctic Tern than to the Common Tern. Arctic Ternslay s•nallerclutches in the Bay of Fundy area than they do in arcticor sub-arcticregions. Clutch-sizefor the speciesaverages undertwo eggs. This is significant,for the speciesseems incapable of Vol.1957XXVIII HAWK•LEY,Breeding Pop,lation o/ ylrctic Terns [89 raisingmore than onechick per pair of adults,at leastat the Machias Seal Island colony. The incubationperiod for individualmarked eggs of the Arctic Tern at Machias Seal Island is 22 days. Not all of the birds delay incubationuntil the clutch is complete. Clutch-sizedoes not affecthatching success but two-eggclutches and one-chickbroods containthe optimumnuinber for fledgingsuccess. The fledgingsuccess {Table IV i for the combinedyears 1947 and 1948 was 34.7 per cent figuredon the total numberof eggs. However,47.6 per cent of the_ breedingpairs succeededin fledginga chick. Non-breedingin the Arcticcaused by late summers,ecological. changes in the ternery, "internal" factorsand predation,reduce populations. Eggingprobably has a seriouseffect on populationsonly in primitive areas. Estimatesbased on average fledging success and banding of essentially all youngready to fledgeplaced populations for the MachiasSeal Island colonyat approximately2,900 pairs for 1947 and 3,450 pairsfor 1948. The populationappeared to be as large or larger in 1949. Suggestionsfor managementinclude control of vegetationand predators,including domestic animals.

LITERATURE Crl'ED ARMSTRONG,EDWARD A. 1954. The behaviourof birds in continuousdaylight. Ibis 96: 1-30. AUSTIN,O. L., SR. 1934. The statusof Cape Cod terns in 1934. Bird-Banding 5: 155-171. 1940. Some aspects of individual distribution in the Cape Cod tern colonies. Bird-Banding 11: 155-169. 1942. The life span of the Common Tern (Sterna hirundo). Bird- Banding 13: 159-176. 1945. The role of longevity in successfulbreeding by the Common Tern. Bird-Banding 16: 21-28. 1946. The status of the Cape Cod terns in 1944; A behaviorstudy. Bird-Banding 17: 10-27. 1948a. Predation by the CommonRat (Rattus norvegicus)in the Cape Cod colonies of nesting terns. Bird-Banding 19: 60-65. 1948b. Status of the Cape Cod terns in 194.7. Bull. Mass. ,4ud. Soc. 32: 137-138. Aus'rIN, O. L., JR. 1929. Contributions to the knowledge of the Cape Cod Sternin•e. Bull. N. E. Bird-Bandi•g ylssoc. 5: 123-150. 1932. Birds of Newfoundland Labrador. Mere. Nuttall Ornithological Club. No. 7. 229 pp. BAILEY,A.M. 1925. A report on the birds of north•estern Alaska and regions adjacent to Bering Strait. Part III. Condor 27: 101-109. -- 1926. A report on the birds of northwesternAlaska and regions adjacent to Bering Strait. Part VIII. Condor 28: 84-86. ---- 1927. Notes on the birds of southwestern Alaska. yluk 44: 1-23. BEa½•AN,G. 1941. Om siInf&glarsoch vadares hackning i lnasfagelkolonier. Dansk Ornithol. Foren. Tidsskr. 35: 120-123. BICkERTON,W. 1909. British nesting terns. Proc. Zool. Soc. Lond., 1909: 800- 802. BraD,C. G. and E.G. 1935. The birds of Jan Mayen Island. Ibis 77: 837-855. -- 1940. Some remarks on non-breedingin the arctic, especially in northeast Greenland. Ibis 82: 671-678. --- 1941. The birds of northeast Greenland. Ibis 83: 118-161. BItACKmLL,HEItV•¾. 1952. Light intensity and waterfowl flights; preflight activities. •ilson Bull. 64.' 242-244. BaOWN, F.A. 1911. Macbias Seal Islands. Bird Lore 13: 239-245. 90] HXW}•LEY, Breedb,g Pop•tlation oj Arctic Terns Bird-BandingApril

BUTURLIN,S.A. 1906. The breeding-groundsoœ the Rosy Gull. Part III. Ibis 48: 661-666. CALL, A. B. 1891. Notes at random. OSlogist 8: 198. CLAY,T. 1948. Relationshipswithin the Sterninae as indicated by their mallo- phagan parasites. Ibis 90: 141-142. 1949a. Species of the genus Saemundssonia (Mallophaga) from the Sterninae. Am. Mus. Novitates, No. 1409: 1-25. 1949b. Some problems in the evolution of a group of ecto parasites. Evolution 3: 279-299. COTT, [t. B. 1949. The paiatability of the eggs of birds. OSlogist Rec. 23: 1-9. 1953-1954. The exploitation of wild birds for their eggs. Ibis 95: 409-449, 673-675; 96: 129-149. CROWELL,E. M., and S. CROWELL.1946. The displacementof terns by Herring Gulls at the Weepecker Islands. Bird-Banding 17: 1-10. DALGETY, C. T. 1936. Notes on birds observed in Greenland and Barfin Island, June-Sept. 1934. Ibis 78: 580-591. DALGLEISH,J.D. 1886. Discovery of the nest of Larus rossii in Greenland. Auk 3: 274. DAaLING,F. F. 1938. Bird flocks and the breeding cycle. Cambridge. Univer- sity Press. x q- 124 pp. DUFFEY, E., and D. E. SERGEANT. 1950. Field notes on the birds of Bear Island. Ibis 92: 554-563. DURANGO,S. 1949. The nesting associationsof birds with social insects and with birds of different species. Ibis 91: 140-143. EIFRIG,C. W.G. 1905. Ornithological results of the Canadian "Neptune" Expe- dition to Hudson Bay and northward. 1903-1904. Auk 22: 233-241. EKLUND,C.R. 1944. Nesting notes on the Arctic Tern. Auk 61: 648. EVANS,E., and W. STURGE. 1859. Notes on the birds of western Spitzbergen, as observed in 1855. Ibis 1: 166-174. EVANS,W. 1891. On the periods occupiedby birds in the incubation of their eggs. Ibis 33: 52-93. FABmctus,E. 1938. Ueber die Abh[ingigkeit der Reiherente, Nyroca juligula (L.), yon den Lariden als Brutvogel im Sch[irmhof. Ornis Fennica 14: 115-125. FIEEDEN,R.A. 1877. List of birds observed in Smith Sound in the polar basin during the arctic expedition of 1875-76. Ibis 19: 401-412. FLOYD, C. B. 1932. Strange disappearance of nesting Penikese Island terns. Bird-Banding 3: 173-174. -- 1938. Terns and tractors. Bull. Mass..dud. Soc. 22: 2-4. FoaBusm E. H. 1921. Report of Edward Howe Forbush, field agent for New England. Bird Lore 23: 345-349. FRIEDMANN,H. 1933. Notes on some birds of Goodnews Bay, Alaska. Condor 35: 239-240. GROSS,A. O. 1937. Birds of the Bowdoin.MacMillan Arctic Expedition 1934. .4uk 54: 12-42. • 1948. Gulls of Muskeget Island. Bull. Mass..dud. Soc. 32: 43-47. HAARTMAN,L. von. 1938. Zur Kenntnis der Brutpsychologie der Reiherente, Nyroca /uligula (L.) Ornis Fennica 14: 125-134. HAW•CSLEY,J. 1950. Notes on Leach's Petrel. Bird-Banding 21: 60. HAWKSELY,0., and J. HAWKSLEY. 1949. Laughing Gull nesting off coast of Maine. Bull. o/ Maine .dud. Soc. 5: 77. Hoe•c•NS,G. H. E. 1949. Stray notes on Mallophaga--IX. /Inn. and Mag. of Nat. Hist. (12) 2: 29-54. HOWELL,J. C. 1948. Observations on certain birds .of the region of Kodiak, Alaska. Auk 65: 352-358. KtmCMAN,F. B. 1908. Variation in the nests of the Arctic and Common Terns. Brit. Birds 2: 78-82, 101-108. KULLENBERG,B. 1946. i2ber Verbreitung und Wanderungen yon Vier Sterna- Arten. Stockholm. /Irkiv /iir Zoologi. Band 38A, No. 17. 80 pp. Lacs:, D. 1933. iNcsting conditions as a factor .controlling.breeding time in birds. Proc. Zool. Soc. Lond. 1933: 231-237. -- 1947. The significanceof clutch-size. Part I. Intraspecific variations. Part II. Factors involved. Ibis 89: 302-352. 'Vol. XX¾III :t957 }IAwKSE¾,Breedi, g Pop,lation o./ Arctic Terns [91

--1948. The significance of clutch-size. Part I1|. Some interspecific comparisons. Ibis 90: 25-45. LINTON,E. 1928. Notes on trematodeparasites of birds. Proc. U.S. Nat. Mus. 73: 1-36. MANNICHE, A. L. V. 1910. The terrestrial mammals and birds of northeast Greenland. Meddelelscr om GrOnland 45: 1-200. MAm'Lr.S, G., and A. MAaPLr.S. 1934. Sea terns or sea swallows. London. Country Life Press. xxi q- 227 pp. MARSHALL,A. J. 1938. Bird and animal activity in the arctic. Jour. Animal Ecol. 7: 248-250. -- 1952. Non-breeding among arctic birds. Ibis 94: 310-333. MaaSHaLL, N. 1942. Night desertion by nesting Common Terns. Wilson Bull. 34: 25-31. McCABr., T. T., and K. RACEY. 1944. The Parasitic Jaeger (Stercorarius parasiticus). Aak 61: 465-466. MONTAGUE,F. A. 1926. Further notes from Spitzbergen. Ibis 68: 136-151. NOI•TON,A.H. 1907. Report of Arthur H. Norton on coloniesof birds in Maine receiving special protection in 1907. Bird Lore 9: 319-327. PABCr.•-W•LKZS,A. H. 1922. On the breeding-habitsof the Turnstone as observed in Spitzbergen. Brit. Birds 15: 172-179. PAL•r.a, R. S. 1941. A behavior study of the Common Tern (Sterna hirundo hirando L.). Proc. Bost. Soc. Nat. Hist. 42, No. 1, 119 pp. PALMGREN,P. 1935. f2ber den tagesrhythmusder viSgelim arktischensommer. Ornis Fennica 12: 107-121. PAYNTEa,R. A., Ja. 1949. Clutch-sizeand the egg and chick mortality of Kent Island Herring Gulls. Ecology 30: 146-166. 1954. Interrelations between clutch-size, brood-size, prefiedging survival, and weight in Kent Island Tree Swallows. Bird-Banding 25: 35-58, 102-110, 136-148. PETtINt;ILL, O. S., Ja. 1936. Impressions of Grand Manan bird life. Wilson Ball. 48:111-119. -- 1939. History of one hundred nests of Arctic Terns. Auk 56: 420-428. PLr.SKE, T. 1928. Birds of the Eurasian tundra. Boston Soc. o.[ Na.t. Hist. Mere. 6 t 3): 107-4.85. RAWCL•FFE,C.P. 1949. Kestrel taking young Arctic Tern. Brit. Birds 42: 90. ROBERTS,B. B. 1934. Notes on the birds of central and southeast Iceland, with special reference to food habits. Ibis 76: 239-263. ROBINSON,H. W. 1920. Disappearanceof nesting species in the Scilly Isles. Brit. Birds 14: 65-66. -- 1921a. Status of Arctic Tern in Lancashire. Brit. Birds 14: 281. ---- 1921b "Status of Arctic Terns ;_n Lancashire and the Farnes" i Letter to editors•. Brit. Birds 15: 92. SALOMONSEN,F. 1947. Maagekolonierne paa Hirsholmene. Dansk Ornithol. Foren. Tidsskr. 41: 174-186. ---- 1955. Bird preservation in Greenland l Fredning af fuglelivet i ArOn- land.• Da,sh Ornithologish Forenings Tidsshri/t 49: 3-11. SELIGMAN,O. R., and J. M. Wt[I•COX. 1940. Some observations on the birds of Jan Mayen. Ibis 82: 464-479. SHELFORD,V. E., and A. C. TWOMEY. 19•1. Tundra animal communities in the vicinity of Churchill, Manitoba. Ecology 22: 47-69. SLATEa, H. H., and T. CAnTEa. 1886. Notes from northern Iceland in the summer of 1885. Ibis 28: 45-52. SUGDEN,J. W. 1947. Exotic eggs in nests of California Gulls. Condor 49: 93-96. SUOMA/AINEN,H. 1939. t)ber die Brutbiologic der Kiistenseeschwalbe,Sterna mactufa Naum., in i,hrem verh•iltnis zu derjenigen der Flussesschwalbe, Stcrna b. birundo L., in den Schi/ren des Finnischen Meerbusens. Ann. Zool. Soc. Zool.-Bot. Fenn. l•anamo 6: 16-23. SUTTON,G. 5I. 1932. The exploration of Southampton Island, Hudson Bay; Part II, Sect. 2. The birds of SouthamptonIsland. Memoirs of the Car- negie Museum. 12. 275 pp. TINUERGEN,N. 1932. Vergelijkende warnemingen aan enkele meeuwan en sterns. Ardca 21: 1-13. 92] •ORTH,HAMPARI4N, R•KE, Survey o./ Ornithosis Bird-Banding April

TOWNSEND,C. W. 1923. Notes on the birds of Grand Manan, New Brunswick. Can. Field Nat. 37: 141-144. TREVOR-BATTYE,A. 1897. The birds of Spitzbergen, as at present determined. Ibis 39: 574-600. VA• Oom)T, G.J. 1934. Ueber einem biolagischenunterschied zeischen Fluss-und Kustenseeschwalbe. Bietr. Fortp/I.-biol. FOg. 10: 5-6. WILLETT, G. 1918. Bird notes from Forrester Island, Alaska. Condor 20: 85. WILLEY, C. H., and H. W. STUNKARD.1942. Studies on pathologyand resistance in terns and dogs infected with the heterophyid trematode, Cryptocotyle lingua. Trans. •4mer. Miscroscop. Soc. 61: 236-253. WILLIAMS, J. G. 1941. On the birds of the Varanger Peninsula, East Finmark. Ibis 83: 245-264. WILLIAMS, R. B. 1947. Notes on the Arctic Tern in Alexander ArchipElago, southeastern Alaska. •4uk 64: 143-144. WILLIAMSON,K. 1948. Arctic Terns scavenging. Brit. Birds 41: 31. WITHERBY, H. F., F. C. R. JOURDAIN,N. F. TICEHURSTand B. W. TuCS:E•. 1941. The handbookof British birds. Vol. 5. London. I4. F. and G. Witherby Ltd. xii q- 356 pp. WmCHT, W. C. 1909. Arctic Terns feeding on crane and May flies. Brit. Birds 3: 91. Central Missouri State College, IVarre•sb•trg, Mo.

A SEROLOGICAL SURVEY OF ORNITHOSIS IN BIRD BANDERS* C. BROOKEWORTH, VINCENT HAMPARIANI' AND GEOFFREYRAKE

INTRODUCTION Ornithosisis a chlamydialt somewhatvirus-like) disease occurring at timesamong birds of manyspecies including pheasants, chickens, pigeons,turkeys, ducks, certain wading birds and evenfuhnars. It is related to, and may be identical with, psittacosiswhich occursin psitmcinebirds. All avian speciestested to date have beenfound to be susceptibleto infectionby experimentalinoculation. There have beenwell.-marked epidemics associated with commercialenterprises such asbreeding aviaries, t.ransport agencies, poultry and pheasant farms and poultry-processingplants. It has occurredalso amongbird huntersin Louisiana. Infection of human beingsby the causativeagent may be followed by mild or severeillness or may not be attendedby clinical symptoms. Little is known of the occurrence of ornithosis in the native wild avifaunaof the United States. Only the Herring Gull, SnowyEgret, Willet, Magpie, Painted Bunting and Goldfinch have thus far been found naturally infected. If the diseaseexists among the hundredsof othernative species to any appreciableextent, especially among other * ContributionNo. 6 from Microbiologyin Medicine. This studyhas been aided by a grant from the Hartz Mountain Products Corporation. ? The work presentedin this paper was performed in partial fulfillment of the requirementsof the Ph.D. degree at the University of Pennsylvania. Present address: Children's Hospital, Philadelphia.