Are wader flocks random grouping?- a knotty problem

BRIAN A. HARRINGTON & LINDA E. LEDDY

Manomet Bird Observatory,Manomet, MA 02345, USA

Citation: Harrington, B.A. & Leddy, L.E. 1982. Are wader flocks random groupings? -a knotty problem. Wader Study Group Bull. 36: 20-21.

Ornithologistsfrequently work on the assumptionthat birds are randomly associated in theirflocks. Our find- ingswith Knots,and those of Furness& Galbraith(1980) with anotherspecies of wader,suggest such assump- tionsmay be incorrect.Our resultof non-randomoccurrence of markedbirds between flocks indicates that careshould be usedin calculatingpopulation estimates using ratios of marked/unmarkedbirds unless broad, representativesampling is achieved.

In a recent Bulletin article, Furness & Galbraith (1980) particular location which normally was heavily used by described a non-random distribution of colour-marked humans. Siesta Key is about 7 km south of SandsPoint wadersand speculated on thepossible causes of theirobser- where the Knots were banded. A count of marked/unmarked vations.We found an analogoussituation during three brief Knotsin one sectionof the SiestaKey flock yielded a ratio surveysof Red KnotsCalidris canutus on the Floridawest of 30/535 (= 5.6%) whereas the ratio in another sec- coastin 1981 and 1982, andspeculate on the significanceof tion of 60/519(= 11.6%)was significantly higher Z 2 = 10.9, our observationswith respectto populationestimates based p < 0.01). Similarly, in one part of a foraging flock on a uponsightings of colour-markedwaders, and with respectto sandbarnext to Siesta Key on 8 Januarythe marked/un- non-random association of Knots within and between flocks. markedratio was 80/739 (10.8%) while it was 16/512 (3.1%) We captured and colour-marked 238 Red Knots (all in anothersection (Z 2 = 23.4,p < 0.01)of thesame flock. The adults)from a flock of 450 restingat SandsPoint, Longboat foragingflock took flight severaltimes when disturbed,yet Key near Sarasota,Florida on 6 January1981 by using a the tendencyfor markedbirds to be concentratedin one sec- rocket net (60 x 40 feet, 1 inch square,knotless nylon net) tion of the flock persisted,even though not necessarilyin the propelledby four rockets.All birds were processedat the same area of the tidal flat. capturesite immediatelyfollowing capture,marked with a saturatedsolution of picnic acid and95% ETOH, and were Non-random distribution between flocks releasedwithin sixhours of thecapture time. The birdswere marked also with colored leg-flags. Subsequently we Not only wasthere non-random distribution of markedKnots censusedflocks of Knots in the Sarasotaregion from 7-9 within flocks,but therewere alsosignificantly different fre- January1981, 9-16 October1981 and 18-21 January1982, quenciesbetween flocks (Table 1). For example,between 7 concentratingour work in placeswhere we knew Knotswere and 10 January1981 the proportionof marked birds (29/ gatheringto rest at high fidesor to feed alongbeach fronts 1059 [2.7%]) found at the banding location (SandsPoint) on falling tides. was significantlylower than the proportion(186/2305 [8.1%] In mostcases it wasnot possibleto searchsystematically foundat SiestaKey, 7 km to thesouth (Z 2 = 32.3,p< 0.01). throughcomplete flocks becausebirds were often milling A non-randomdistribution persisted 8 and 12 monthslater about,or in somecases were repeatedly flushed by pedestri- (Table 1), after all the birds had completeda pre-alternate ans. Our aim in searches was to obtain an estimate of the andpre-basic molt, anda round-tripmigration to their arctic flocksize and a' tally of theproportion of themarked birds breedingrange. present.In somecases we wereable to searchthrough a flock more than once, resultingin a tally of birds checkedthat Discussion exceededthe flock size, while in othercases it was not pos- sible to check a whole flock. The results of non-random distribution of marked Knots have Whateverthe situation,we alwaystried to collectratio from two importantimplications for waderresearch. First, the non- representativesections throughout the flocks.It was always random distributionof marked birds within flocks suggest possiblethat marked birds could have been counted more than that flocks are not random associations,a curious and unex- one time or that some marked birds in flocks were not seen. pectedfinding. Exceptas described,the markedbirds were not groupedin any way we could see,but insteadwere dis- Results tributedamongst unmarked birds. We saw no evidenceof segregationaccording to the presenceor absenceof dye and/ Intra-flock distribution of marked birds or bands,no unusualaggressive behaviour, no evidenceof fidelity by markedbirds to particularspots, no evidenceof While countingKnots we noticedthat marked birds were not territorialbehaviour, nor did we seeany otherunusual reac- randomlydistributed in eitherforaging or restingflocks. For tionto markedbirds by theirneighbours. We areleft with the example,on 7 January(Table 1) a large restingflock was idea that Knot flocks are not random associationsof birds, gatheredduring a stormon the upperbeach at the northend and that there is somesort of socialorganisation to flocks. of SiestaKey, the only time we found Knots restingat this In contrast,Pete Myers (pers.comm.) did not find thatSand- 51 Bulletin69SpecialIssue July1993 52 Wader Study Group Bulletin

Table 1, Frequenciesof Red Knotsbanded at Sands Point on 6 January 1981 and subsequentlyfound in flocks at various locations.

Location Distance (km) Flock size No. of birds No. marked Percent marked from banding site checked

7 January1981 SiestaKey 7 4,200 1,054 90 8.5% 8 January 1981 SandsPoint 0 400 400 10 2.5% 8 January 1981 SiestaKey 7 2,230 1,251 96 7.7% 9 January 1981 SandsPoint 0 900 659 19 2.9% 7 Maderia Beach 62 550 622 15 1.4% 9 October 1981 Sands Point 0 350 735 31 4.2% 16 October 1981 Indian Shores 69 800 1,130 9 0.8% 18 January 1982 ManasotaKey 42 1,125 1,906 29 1.5% 21 January1982 LongboatKey 15 1,499 781 31 3.9%

erlings Calidris alba in California had complex, within- number we estimated(3,875) from countsmade in an aerial flock, socialorganisation. survey on 31 . Second, our result of non-random occurrence of marked To summarize,ornithologists frequently work on the birds between flocks indicates that care should be used in cal- assumptionthat birds are randomly associated in theirflocks. culating population estimates using ratios of marked/ Our findings with Knots, and thoseof Furness& Galbraith unmarkedbirds unlessbroad, representativesampling is (1980)with anotherspecies of wader,suggest such assump- achieved.For example,at SandsPoint in January1981 we tions may be incorrect. found 29 of the 238 marked Knots among 1,059 birds checked,giving an estimateof 8,691 Knotsin the areapopu- Reference lation. On the otherhand, at SiestaKey we founda ratio of 186 markedto 2,305 unmarkedKnots, leading to population Fumess, R.W. & Galbraith, H. 1980. Non-random distribution in roost- estimateof 2,949 Knots. Combining the ratios from both ing flocksof wadersmarked in a cannonnet catch. Wader Study Group Bull. 29: 22-23. areasgives a populationestimate of 3,724,much closer to the

Rationale and suggestions for a hemispheric colour-marking scheme for shorebirds: a way to avoid chaos

J.P. MYERS•, J.C. MARON2, E. ORTIZ3, G.V. CASTRO3, M.A. HOWE4, R.I.G. MORRISON 5 & B.A. HARRINGTON 6

•Academyof NaturalSciences, 19th and theParkway, Philadelphia, PA 19103,USA; 2Dept of Biology,Univer- sityof NorthDakota, Grand Forks, ND 58201, USA;3APECO, Atahualpa 335, Lima,; 4patuxent Wildlife ResearchStation, Laurel, MD 20708, USA; 5CanadianWildlife Service, 1725 WoodwardDrive, Ottawa, Canada,KIA OE7; 6ManometBird Observatory, Manomet, MA 02345, USA

Citation: Myers, J.P., Maron, J.C., Ortiz, E., Castro, G.V., Howe, M.A., Morrison, R.I.G., & Harrington, B.A. 1983. Rationale and suggestions for a hemispheric colour-marking scheme for shorebirds: a way to avoid chaos. Wader Study Group Bull. 38: 30-32.

Within the next few years the number of shorebirdcolour systemis complex,but so is the problemit addresses.No marking projectsin the New World is likely to expand perfectsolution is availablegiven the diversityof interests considerably.One commongoal sharedby many of these and countries and the limitations on materials and resources. programsis trackingmovements between breeding and win- We hopethis systemstrikes a reasonablebalance. We hope teringgrounds, and identifying stop-over sites used by spe- thatresearchers will providefeedback to us on its usefulness cific populationsduring migration. The prospectfor success and drawbacks. in thiswork is enhancedgreatly by expandingthe numberof projectsinvolved. But this expansionmay resultlargely in 1. This systemuses coloured leg bandsand flags made of a confusion, unless considerable effort is made to co-ordinate plastic,darvic, with UV-stablecolours to preventfading. The colour-marking schemes.We outline here a systemthat flagsprovide extraordinary visibility when placed beneath shouldfunction if adoptedby shorebirdbanders working in the tarsometatarsaljoint of shorebirds.The flags are long- the New World. lived andcolour-fast, and until 1982 were virtuallyunused The systemwe proposedis basedon a seriesof practical in New Word shorebirdbanding. Flags are immediately dis- considerationsand researchgoals, summarisedbelow. The tinguishedin the field from colouredleg bands.Thus this • Bulletin69SpecialIssue July1993