DIFFERENCES IN SUSCEPTIBILITY TO PREDATION DURING NESTING BETWEEN PIED AND BLACK (HIMANTOPUS SPP.)

RAYMOND J. PIERCE1 Departmentof Zoology,University of Otago,Dunedin,

ABSTRACT.--Thenesting success of the Pied (Himantopushimantopus leucocephalus) and the endangeredBlack Stilt (H. novaezealandiae)was studiedfor three consecutivebreeding seasons(1977-1979) in New Zealand. Black Stilts had a breeding successof lessthan t%, comparedwith over 8% for Pied Stilts. Predation by fetal mammalswas the main causeof nest failure for both species,but the impact was greater on Black Stilts. Direct and indirect evidence (including trapping predators)suggests that these introduced ground predators kill manychicks. Several factors made Black Stilts vulnerable, particularly their nestingalong streambanks that were often frequentedby predators;Pied Stiltsnested in swampswhere predatorswere few. BlackStilt chickstook up to 2 weekslonger to fledgeand their foraging patternsmade them more vulnerablethan chicks.Other factorsthat possibly increasedpredation risk of BlackStilts include nestingat times of high predatoractivity, solitarynesting, high site fidelity, ineffectivedistraction displays, and lack of a disruptive camouflagepattern in adults.Introduced ground predators probably have contributed great- ly to the declineof BlackStilts. Pied Stiltshave not beenaffected similarly, because their evolutionarypast has included long exposureto groundpredators. Received 27 February1984, acceptedI July 1985.

Two speciesof stilts () occur data for most speciesare lacking. In this study in New Zealand: the endemic (Hi- of stilt ecology in South Canterbury, I exam- mantopusnovaezealandiae) and the Pied Stilt (H. ined the nesting successof Pied and Black stilts himantopusleucocephalus), which is the Austral- and comparedtheir susceptibilityto predation. asianrace of the widespreadBlack-winged Stilt. The BlackStilt probably has occurredin New STUDY AREA AND METHODS Zealand for many thousandsof years,but the Pied Stilt is a more recentimmigrant, probably Mostdata were collectedduring three field seasons having arrived in the late 18th or early 19th (1977-1978, 1978-1979, and 1979-1980) in the lower century (Fleming 1962, Pierce 1984a).During 15 km of the CassRiver Valley, LakeTekapo at 710- the 19th century Black Stilts were widespread 850 m (Fig. t). in New Zealand riverbeds and nested at least The study site is a glaciated valley flanked by as far north as central North Island, but they mountains,with tussocks(Festuca spp., Poaspp., and declined rapidly. Since the 1950's a small Chionochloaspp.) and other grassesand small shrubs breeding population has remained only in cen- on the lower slopes.Stilts of both speciesnested in four habitats: riverbed, side streams, ponds, and tral (Pierce 1984a). By contrast, swamps.The riverbed consistedof bare shingle and Pied Stiltshave expandedrecently and are now had braided channelswith widely varying flows common in most types of wetland throughout (mean annual flow was about t0 m•/s). The side New Zealand. streamshad relatively constantflows, and a variety Stilts are ground nesters and are therefore of grassesand other small plants grew on theirbanks. vulnerable to predation by mammals. The in- Muddy ponds were surroundedby grasslandand troduction of carnivorous mammals in the 19th usually containedwater all year. The swampswere century often is claimed to have had a delete- small (

273 The Auk 103:273-280. April 1986 274 RAYMONDJ. PIERCE [Auk,Vol. 103

TABLE1. Breeding successof Pied and Black stilts.

Black Stilt

Unpro- Pro- Pied Stilt tected tected

Total nests 125 27 23 Nest days (N) 1,374 305.5 353 aodley % Total failed nests(F) 49 19 12 Total eggshatched a 260 33 40 CassLake • Total fledged young 69 2 13 RiverTekapo Probability of nest surviving to hatchingb (A) 0.349 0.155 0.367 Probability of egg hatching in a sur- viving nest• (B) 0.903 0.971 0.909 Probability of chick surviving to fledg- ing (C) 0.265 0.061 0.325 Breeding successd 0.084 0.009 0.108

Wamonc ' Mean clutch size of Pied Stilts was 3.8 (range 3- 4); mean clutchsize of BlackStilts was 4.0 (range 3- 6). b(1 - F/N) 29.The duration of egg-layingand in- cubationwas 29 days for both species. "'•• cThe remaining eggswere infertile or addled. a Probabilityof egg producingflying young (= A x Fig. 1. Location of study areas. The main study BxC). area was in the CassRiver Valley.

On every sixth day of the breeding seasonI walked Wildlife Service. I found no evidence of mammals the wetland nesting grounds to check whether in- scavengingon desertedstilt eggs.Deserted eggs of dividual pairs were laying, incubating, or guarding stilts (3 clutches)and of other charadriiform species chicks. Other data collected were habitat, substrate, were usually still intact two weeks after having been distanceto water (all after the eggshad hatched),and abandoned. proximityof other nestingstilts. Inevitably somenests To test the hypothesisthat breeding successwould would have been abandoned before I found them, so increaseif predator numberswere reducednear nests, I used the Mayfield (1975) method for calculating I set traps around 23 randomly selectedBlack Stilt breeding success.Because nest checkswere frequent, territories. Gin or Fenn traps were set at intervals of I used the midpoint assumptionrather than the May- 40-90 m (average 60 m) and no closer than 30 m to field -40% method (Miller and Johnson 1978, John- nests.Each trap was placed under a Discariabush or son 1979). Failed nests were examined for evidence covered with a wooden tunnel 30-45 cm long x 20- of the causeof failure. I identified nest predators as 24 cm wide x 16-20 cm high. The traps were baited (1) (Mustelafuro), 13 nestsin which the eggs with rabbit flesh and checkedevery secondday. The were taken whole (no eggshellleft), 10 of which had successof these "protected" Black Stilt nests was ferret footprints leading to or from the nest and 4 of comparedwith 27 Black Stilt nests without traps which had ferret faeces at the nest; (2) feral cat (Felis ("unprotected").In casepredators followed my scent catus),11 nests in which 4-6 large and many small to nests (Bart 1977), I visited only those nests that eggshell fragments were in or near the nest, 6 of could be approached through water. When I was which had catfootprints and 5 of which had catfaeces within 10 m of these nests I recorded the reactions nearby;(3) Norway rat (Rattusnorvegicus), 14 nestsin of the adult to my presence,as well as their which very small shell fragmentswere scatteredin nesting stage. Nocturnal observationsof breeding and around the nest, 3 of which had rat footprints behavior were made using an NVC night vision sys- and 2 of which had rat faecesnearby; or (4) Austra- tem with 500-mm lens from a hide or tent 60-200 m lasian Harrier (Circusapproximans), 2 nests in which away. more than half of each eggshellremained in or near To establish chick survival, observations of brood the nest, eachshell with one or two puncturemarks; size were made with a telescopeabout every 10 days 1 nest also had a harrier pellet present. I identified for 9 protectedBlack Stilt broods'and 6 unprotected mammalsigns from the guide of Lawrenceand Brown Black Stilt nests over 3 yr. I abandonedsimilar at- (1973) or with help from staff of the New Zealand temptswith Pied Stilt chicks,because family groups April1986] PredationonStilts 275 were usually closely spacedand therefore easy to T^BLE 2. Cause of nest failure of Pied and Black stilts. confuse. It was not until juvenile Pied Stilts were flying (and lesswary) that family sizescould be es- Side tablished. Adult Pied Stilts were individually rec- stream River Pond Swamp Total ognizable all seasonby their plumage markings Pied Stilt (Pierce 1984b),and many were also banded with in- Total nests 16 24 15 70 125 dividual color combinations. Family parties with Preyedon 5 2 6 11 24 flying youngusually stayed near the nestingarea for Flooded 0 13 0 3 16 a few days at least. To be certain that the correct Other a 3 2 0 4 9 number of young was established,however, I made Black Stilt unprotected more visits than usual to pairs about to fledgeyoung. Total nests 15 2 3 7 27 Preyedon 7 0 3 1 11 Flooded 1 0 0 0 1 Other a 4 1 0 2 7 RESULTS Black Stilt protected BREEDING SUCCESS AND CAUSES OF FAILURE Total nests 13 0 6 4 23 Preyed on 3 0 2 0 5 The breeding successof BlackStilts was much Flooded 3 0 0 0 3 Other a 2 0 1 1 4 lower than that of Pied Stilts (Table I). The probability that an egg present at the begin- a"Other" includesdesertion, damage by stockand ning of incubation would produce a fledgling wind, and unknown causes, some of which could have been predation. was 0.084 for Pied Stilts and 0.009 for unpro- tectedBlack Stilts. Eggsin protectedBlack Stilt nests had a much higher probability of pro- and chick mortality. I had no direct evidence ducing a fledgling (0.108) than those in unpro- of predatorstaking Pied Stilt chicksbut found tected nests. two cases in which Black Stilt chicks were tak- Predation by mammalsand birds accounted en by a ferret and a cat. for 49% of Pied Stilt failures and 64% of Black Indirect evidence of the impact of predators Stilt failures for which the cause was estab- on BlackStilt chickscame from using two pred- lished (Table 2). The impact of predation was ator-proofexclosures (Pierce 1982). Before these more severe on Black Stilts. At least 41% of the sites were fenced, four nesting attempts by unprotected nests were preyed upon, com- Black Stilts were unsuccessful. Mammals took pared with only 19% in Pied Stilts (X2 = 7.37, eggsand chicksin 1977,1978, and 1979.Of the P < 0.01). Even protectedBlack Stilt nestswere 30 chicks that have hatched in these exclosures preyedon, but proportionatelyfewer (22%)than since 1980, about 60% have flown. This high for unprotected nests. , feral cats, Nor- productivity contrastswith a fledging rate of way rats, and harriers all ate eggs (Table 3). only 7% at unprotectedBlack Stilt nests that Three other potential predators in the study hatched eggs (Table I). area were (Mustela erminea) and (M. nivalis), both of which were uncommon, POSTBREEDING MORTALITY and hedgehogs (Erinaceuseuropaeus), which were common. Thirteen of 15 young Black Stilts survived Exceptin severe conditions,Black Stilt nests until the end of winter (August), when the were not very susceptibleto flooding because family units split up. This representsa mortal- the birds tended to nest in stable situations ity rate for flying young of 13%for 6 months, where food density was high. or 0.5% per week. No nests were known to be deserted after The longevity of Black and Pied stilts is un- light snowfalls (up to I0 cm of snow), but in known, but an adult female Black Stilt (2+ yr late October 1982, a heavy snowfall of 25-35 old) banded at its nest in December 1973 (R. J. cm caused almost all Pied and Black stilts to Nilsson pers. comm.) was still alive and nesting desert their eggs. in November 1983at the age of over 12 yr. This Becausestilts are nidifugous, it is very diffi- is probablyan exception,however, as sev- cult to establisheven by indirect means how eral banded Black Stilts disappearedat an age chicksdie. Exceptduring two heavy snowfalls, of 2-3 yr, indicating that the potential life span there was no correlation between bad weather may rarely be realized.Of I0 color-bandedBlack 276 RAYMONDJ. PIERCE [Auk,Vol. 103

TABLE 3. Predation on Pied and Black stilts in the CassValley, 1977-1979.Figures in parenthesesare from the Godley Valley. leucocephalus 60' Nor- 1977-79 Fer- Fetal way Har- Un- ret cat rat tier known Total 40' Pied Stilt Adult 0 1 0 0 1 2 Eggs 4 5 12 2 1 24 Chicks 0 0 0 0 0 0 2O Black Stilt Adult 0 2 0 0 1 3 Eggs 8 (1) 4 (2) 2 0 6 (2) 20 (5) Chicks 1 1 0 0 0 2 , B. o novaezealandiae Stilt fledglings in 1977-1979, only 4 survived c; %%1977-79 to 3 yr of age and 1 to 4 yr. Nesting was a vulnerable period for adult z 2o •,• //1970-80 Black Stilts, with 5 of 7 known deaths occur- ring then. Three of thesebirds were taken with their eggsby mammals.Two birds disappeared during the fledging period of their young in 1980, but the cause of death was not estab- C. lished. In two casesfollowing the death of one member of a pair, the nestingarea was not used 2 R• R= Norway Rat for nesting by Black Stilts again, despite un- • C I C= Feral Cat changedfood suppliesand nest sites.

SOURCES OF BLACK STILT VULNERABILITY E I R • F=Ferret Nestinghabitat.--In the CassValley Pied Stilts O 0 F nestedmainly in swamps,whereas Black Stilts •. C F preferred the banks of side streams. Overall Stream Pond River Swamp choicewas significantly different between the Fig. 2. Nesting habitatsof Pied (A) and Black (B) species(x 2 = 52.29, 3 df, P < 0.001). Few pred- stilts, and the trapping frequenciesof predators in ators were trapped in the swamps,but many those habitats (C). Capture rates of predators are ex- were trappedat the side streams(Fig. 2). This pressedas numbercaught per 100 trap nights,where concentrationof predatorsexerted a high pre- "number of trap nights" is defined as "number of dadon pressure on all stilts nesting at side traps x number of nights of trapping." streams: 31% of Pied Stilt nests and 47% of un- protected Black Stilt nests found at the streams reactions(4 by Pied Stilts, 6 by BlackStilts) to were preyed on (Table 2). The effectsof inter- ferrets and cats followed similar trends. habitatdifferences in prey availability on chick Colonialvs. solitarynesting.--In the CassVal- survivorshipare not known. ley Pied Stilts nestedin loosegroups or small Antipredatorbehavior.--When I was within 10 (averageof 5 nests)colonies and 117 of the 125 m of a Pied Stilt nest, the adults usually per- nestswere less than 100 m (average 18 m, n = formed distractiondisplays, often severalbirds 55) from nests of other Pied Stilts. The other 8 together.The birds displayedat almostall nests nestswere of "solitary"pairs. In contrast,Black being incubatedand also at some during the Stilts nested more than 100 m from other stilts laying period (Fig. 3). Black Stilts often used (Black or Pied) on 20 of 27 occasions. At Pied the aggressiveflight (dive-bombing) method Stilt coloniesthe first bird to detect a predator and less often distraction displays, but there would fly into the air, gaining height quickly was considerablevariation among pairs. Ten and often uttering alarm calls. This behavior April1986] PredationonStilts 277

leucocephalus leucoce •halus 20- 108 166 100]34 103

50.

1C

novaezealandiae

(:• I I •- novaezealandiae

D Month Fig. 4. Nest initiation by Pied (top) and Black Laying Early •••]Late Fledgling (bottom)stilts in the'Cass 0Valley. N incubation incubation period

Fig. 3. Reaction of nesting stilts to a human in- truder. Pied Stilts(top) usuallyperformed distraction of successearly and late in the season(Fig. 5A); displays (open bars) such as false brooding and late-nesting birds had to contend with in- feigning injury; Black Stilts (bottom) usually flew creasedpredator densities (Fig. 5B) resulting about (shaded bars) or flew aggressivelyat the in- from influxes of adult predatorsand especially truder (black bars). kittens. Chickbehavior.--During the night, especially immediately alerted all other birds, and they on calm, moonlit nights, 4-7-week-old Black groupedtogether to attackthe predatoror lure Stilt chicks foraged up to 100 m apart and up it away with distractiondisplays. Black Stilts, to 150 m from the guarding parent. Pied Stilt being mostly solitary, seldomhad this advan- chicksforaged and calledat night also,but they tage. seldomventured even 40-50 m from their par- Of 14 individually known pairsof Pied Stilts ents. The wide-ranging activity of Black Stilt that renested, only 5 did so within 100 m of chicksprobably made them more vulnerable, the old site. Of 15 pairs of renesting Black although I never saw predationactually occur. Stilts, however, 10 renested within 100 m of Fledgingperiod.--The fledging period of 17 the old site. Pied Stilts (œ= 34 days, range 30-37) was sig- Timingof and durationof nesting.--BlackStilts nificantly shorter than for 14 Black Stilts (• = began nesting earlier than Pied Stilts (Fig. 4). 46 days, range 39-55; P < 0.001; all data re- In all three yearsno first nests(mid-September corded between November and January). No to early October)of Black Stilts were success- between-habitatdifferences in fledging period ful, and only one survived to hatching. This were found. The differences were a result of high level of predationmay have resultedfrom slower growth ratesin BlackStilts (Fig. 6). The a shortageof alternative food for predators(R. onset of rapid growth of the chick occurredat Pierce unpubl. data). Pied Stilts had a low rate about 2 weeks in Pied Stilts, but not until after 278 R•YMONDJ.PIERCE [Auk,Vol. 103

150- Weight r=0-94 • 100-a= 9'76 .://f'.- v b=O'09..P.•.'..- .'y. '• './•'..•B . ß •,,•. b=O.06

0 T'arsus s Time of nest initiation E r: 0.93 • a: 25.95 •oSuccessful 0 33 51 17 nests per mo•th = b=O.03 •• 40. • .• • a=-0.97 28.33 -- - :•t2• '• -.' b=O'02 20 150- Wing

a=10'69 '/ • , / 19•.79--•77-78 • ]oo-.b:0.08 ß/..- z•B

0 S • 0 N ' D J F Month ] .... a:..25 Fig. 5. Success of Pied Stilt nests in relation to time of initiation (A) and the number of predators • •. ,,,•:z:•:.... b:0'05 0 ]0 20 30 40 •0 trapped (B). Days after hatching about 3 weeks in Black Stilts. During the very •. 6. Growth o• sdtt ch•cks.gach point (•ed warm 1983 seasonchicks of both speciesgrew Sdtt) a•d c•oss (Btack Sdtt) •ep•ese•ts o•e ch•ck o• more rapidly than usual, with some Black Stilt k•ow• a•e. The expo•e•dat cu•e y = ae• was o• young flying at 35-37 days. best Qt, p•obabty because•owth com•ued beyond Most deaths of Black Stilt chicks occurred in the date o• Q•st Q•ht, affe• which •o measurements coutd be takem • = 2Jed Sdtt, B: Btack Stitt. the first 2 weeksof the fledgingperiod (Fig. 7), but several Black Stilt chicksdisappeared when they were over 5 weeks old, an age at which other shorebirds (Boyd 1962, Hale 1980). most Pied Stilt chickshad fledged. Breedingsuccess for unprotectedBlack Stilts is exceptionallylow, with predationby mammals DISCUSSION being the main cause.The breeding successof other recurvirostridsis considerablyhigher Before carnivorous mammals were intro- than that for Black Stilts (Lippens et al. 1966, duced to New Zealand, mainly in the 19th cen- Cadburyand Olney 1978).The low annual pro- tury (Thompson 1922, Wodzicki 1950), Black duction of Black Stilts, together with the exis- Stilts had few major predators. Native birds tence of much unoccupiedhabitat (Pierce 1982), known to prey on the eggs or young of other suggeststhat predation has contributed consid- birds are the (Gallirallus australis),?ukeko erably to the overall population decline. In ad- (Porphyrioporphyrio melanotus), Australasian dition, low population recruitment facilitates Harrier, and gulls (Larusspp.), while the New the formation of mixed pair bonds (Pied x Zealand Falcon (Falco novaeseelandiae)and har- Black Stilt) with subsequent hybridization rier sometimesprey on stilt-size adult birds (Pierce 1984b). (Oliver 1955, Baker-Gabb 1981). Several features of its biology suggestthat The breeding successof Pied Stilts approxi- the Black Stilt has not developed antipredator matesthe breeding successreported for several behaviortoward mammals.Dive-bombing, an April1986] PredationonStilts 279

30- ß

20-

o

o o

o o o o o o

Weeks after hatching

Fig. ?. Survivorshipof protected(dosed circles) and unprotected(open circles)Black Stilt chicks.

effective deterrent of avian predators (Kruuk Fig. 8. Incubating stilts, showing the disruptive 1964, Sordahl 1981), was used frequently by camouflagepattern of a Pied Stilt (top) and the more Black Stilts but was infrequently used by Pied conspicuousBlack Stilt plumage(bottom). Stilts. The Black Stilt showsstrong nest-sitefi- delity, rather than shifting to a new nest site shorebirds,nested close to "bold" species,they after nest loss (e.g. see Furrer 1979), and this suffered less from predation than did isolated probably increasesthe chancesof double fail- breeders. ure. High chick-adult distancesin Black Stilt With Black Stilts, their side-stream nesting families probably reduce the energy cost of habitat, high site fidelity, solitary nesting, poor vigilance by the parents(e.g. Walters 1982)but distractiondisplays, high chick-adultdistances, at the sametime increasethe risk of chick pre- and slow growth rates in chicks all increase dation. their susceptibilityto mammalian predation. Wide spacing between nests is generally The conspicuousblack plumage itself (Fig. 8) considered to be an adaptation against preda- may assistpredators in locating nestsor young. tion (Tinbergen et al. 1967, Page et al. 1983). It These features probably reflect the absenceof is likely that spacing of Pied Stilt nests was predatory mammalsfrom New Zealand until sufficientlywide that a predator would not be the 19th century. Nesting on the dry banks of attractedto the area, yet also sufficiently dose streamsand ponds,having a high site fidelity, for group distractiondisplays to be performed. wide-roamingof chicks,and using dive-bomb- Solitary-nestingstilts (mostly BlackStilts) were ing on harrierswould have imposedlittle pre- dependent entirely on their own ability to de- dation risk. The Pied Stilt exhibits alternative tect and repel predators,although in somecases featuresin its breeding biology, probably be- Banded Dotterels or other speciesmay have causeof a different evolutionary past that in- provided early warnings of approachingpred- volvedexposure to manytypes of nativeground ators. Eight solitary-nestingpairs of Pied Stilts predatorsin .In New Zealandthe Pied had a breeding successof only 5%, similar to Stilt retained many of its predator-avoidance the successrate of Black Stilts. Conversely, 6 features, enabling it to deal with the intro- Black Stilt nests in or near Pied Stilt colonies duced carnivorous mammals. had an average successof 8% compared with about 1% for unprotectedsolitary nests.These ACKNOWLEDGMENTS differences, however, were not significant. Goranssonet al. (1975) and Dyrcz et al. (1981) My thanks go to Grant Gillespie, Bartie Heather, found that, when "timid" species,including Phil Moors, Mick Clout, and an anonymousreferee 280 R•¾MONDJ.PIERCE [Auk,Vol. 103 for useful discussionand improving earlier drafts of LIPPENS,L., P. MAES, & H. VOET. 1966. De stelklu- this paper; Richard Clarke and Robin Baldwin for teninvasie Himantopushimantopus. Gerfaut 56: assistingwith fieldwork; Kaj Westerskovand Caro- 135-161. lyn Burnsfor thesissupervision; Jim Murray for per- MAYFIELD,H. 1975. Suggestionsfor calculatingnest mitting me to work and camp on his land; the Uni- success. Wilson Bull. 87: 456-466. versity Grants Committee, Royal Forest and Bird MILLER,H. W., & D. H. JOHNSON.1978. Interpreting Protection Society, Acclimatisation Societiesof New the resultsof nestingstudies. J. Wildl. Mgmt. 42: Zealand,and the Royal Societyof New Zealand for 471-476. financial support; and Rose Luxford and Frances MooRs, P. J. 1983. Predation by mustelidsand ro- Wood for typing the manuscript. dents on the eggs and chicks of native and in- troduced birds in Kowhai Bush, New Zealand. LITERATURE CITED Ibis 125: 137-154. OLIVER, W. R.B. 1955. New Zealand birds, 2nd ed. BAKER-GABB,D.J. 1981. The diet of the Australasian Wellington, New Zealand, Reed. Harrier (Circusapproximans) in the Manawatu- PAGE, G. W., L. E. STENZEL,D. W. WINKLER, & C. W. Rangitikeisand country, New Zealand.Notornis SWARTH. 1983. Spacing out at Mono Lake: 28: 241-254. breeding success,nest density, and predation in BART,J. 1977. Impact of human visitations on avian the Snowy Plover. Auk 100: 13-24. nesting success.Living Bird 16: 187-192. PIERCE,R. J. 1982. A comparativeecological study BOYD,H. 1962. Mortality and fertility of European of Pied and Blackstilts in South Canterbury.Un- Charadrii. Ibis 104: 368-387. published Ph.D. dissertation, Dunedin, New CADBtJR¾,C. J., & P. J. S. OLNEY. 1978. pop- Zealand, Univ. Otago. ulation dynamicsin England. Brit. Birds 71: 102- 1983. Charadriiformsof a high countryriv- 121. er valley. Notornis 30: 169-185. DYRCZ, A., J. WITKOWSKI, & J. OKULEWICZ. 1981. 1984a. The changeddistribution of stilts in Nesting of "timid" in the vicinity of New Zealand. Notornis 31: 7-18. "bold" ones as an antipredator adaptation. Ibis 1984b. Plumage,morphology and hybridi- 123: 542-545. sation of New Zealand stilts (Himantopusspp.). FLEMING,C.A. 1962. New Zealandbiogeography: a Notornis 31: 106-130. paleontologist'sapproach. Tuatara 10: 53-108. SORDAHL,T.A. 1981. Predator-mobbing behaviour FURRER,R. K. 1979. Shifting breeding locationafter in the shorebirds of North America. Bull. nest lossin the colonial Fieldfare (Turduspilaris). Study Group 31: 41-44. J. Ornithol. 120: 86-93. THOMPSON, G. M. 1922. The naturalization of ani- GORANSSON,G., J. KARLSSON,S. G. NILSSON,& S. ULF- mals and plants in New Zealand. Cambridge, STRAND. 1975. Predation on birds' nests in re- England, Cambridge Univ. Press. lation to antipredator aggressionand nest den- TINBERGEN,N., M. IMPEKOVEN, & D. FRANCK. 1967. sity: an experimentalstudy. Oikos 26: 117-120. An experimentin spacingout as a defenseagainst HALE, W. G. 1980. Waders. London, Collins. predation. Behaviour 28: 307-321. JOHNSON,D. H. 1979. Estimating nest success:the WALTERS,J.R. 1982. Parental behaviour in lapwings Mayfield methodand an alternative.Auk 96:651- (Charadriidae)and its relationshipswith clutch 661. sizeand matingsystems. Evolution 36: 1030-1040. KRtJtJK,H. 1964. Predatorsandanti-predatorbehav- WODZICKI, K.A. 1950. Introduced mammals of New iour of the Black-headed Gull, Larus ridibundus L. Zealand: an ecological and economic survey. Behav.Suppl. 2: 1-130. Wellington, Dept. Scientificand Industrial Res. LAWRENCE,M. J., & R. W. BROWN. 1973. Mammals Bull. No. 98. of Britain: their tracks,trails and signs.London, Blandford Press.