The Condor97:92C-934 0 The Cooper Ornithological Society 1995

PATTERNS OF HABITAT USE FOR PIED AND SOOTY NESTING AT THE FURNEAUX , ’

BROOK LAURO~ AND ERICA NOL Department of Biology, Queens’ University,Kingston, Ontario, Canada, K7L 3N6 and Biology Department, Trent University,Peterborough, Ontario, Canada, K9J 7B8

Abstract. We studied patterns of habitat use at the levels of generalhabitats (islandsand substrates),feeding territories, and nest sites for two speciesof oystercatchersnesting sym- patrically at the Fumeaux Islands, Australia. Sooty Oystercatchers(Haematopus jiiliginosus) nested only on small islands while Pied Oystercatchers(H. Zongirostris)nested on Fhnders , a large island as well as small islands. Sooty Oystercatchersnested more commonly at rocky shoreswhile Pied Oystercatchersnested more commonly at sandy shores.This was in part related to their selection of islands since small islands were rockier than Blinders Island and to their selectionof foraginghabitats relative to nestingareas. Pied Oystercatchers only placednests on beacheswhere intertidal mudflats were available for foragingterritories; these beaches were generally sandy. Sooty Oystercatchersused either muddy and rocky substrates(the primary available substrates)for foragingand their choiceof feedingterritories played little role in their selection of nesting substrates.At the light colored beachesof the Fumeaux Islands, Sooty Oystercatchers(uniformly black) placed eggsin low visibility nest sitesnext to and under vegetation,while Pied Oystercatchers,with their countershadedcolor pattern, chose higher visibility nest sites on open beach. We suggestthat at this study site Sooty Oystercatcherswere lesscryptic than Pied Oystercatchersand compensatedby choos- ing vegetative characteristicsat nest sites that hid the presenceof their nests from diurnal, visually hunting predators. were the most likely predators to influence choice of vegetative characteristicsat nest sites. Kev words: Ovstercatcher:nest site: feeding territory; island; color pattern: predation; ” visibkty.

INTRODUCTION around the world (Baja California, southern South The oystercatchers form a monogeneric family America, New Zealand and Australia) black and of shorebird (Haematopodidae) whose species pied species of oystercatchers occur sympatri- nest along the coastlines of the world. Black spe- tally (Hayman et al. 1986), usually where there cies of oystercatchers generally nest at small is- is a mixture of rocky and sandy shoreline. Thus, lands where beaches are rocky and dark in color the differences in habitat use for pied and black while pied species of oystercatchers usually nest species of may result in ecological at mainland or island locations on sand beaches segregation, facilitating coexistence. that are light in color (Bent 1929, Heppleston Important factors that may influence patterns 1972, Baker 1974a, Hartwick 1974, Hockey 1982, of habitat use for the oystercatchers include: spe- Hockey and Underhill 1984, Lane 1987, Lauro cies differences in foraging ecology and response and Burger 1989, No1 1989, Lauro and No1 1993). to potential predators given species color pattern. Most species of oystercatchers place nests on the Black species appear to be better adapted to cap- ground above the high tide line in high visibility turing prey on rocky shores (Bent 1929; Baker locations (Heppleston 1972, Hartwick 1974, 1974a, 1974b; Hartwick 1976; Frank 1982; Hockey and Underhill 1984, Lauro and Burger Hockey and Branch 1984; Hockey and Underhill 1989, Lauro and No1 1993). At several regions 1984; Lindberg et al. 1987; Lauro and No1 1995) while pied species of oystercatchers appear to be better adapted for capturing prey on soft sub- ’ Received 26 September 1994. Accepted 22 June strates such as intertidal mudflat (Bent 1929; 1995. Heppleston 1972; Dare and Mercer 1973; Swen- 2 Presentaddress: Department of Computer Science, Mathematics and Science, St. Vincent’s College, St, nen et al. 1983; Hulscher 1985; Baker 1974a, John’s University, Jamaica, NY 11439. 1974b; Goss Custard et al. 1992; Lauro and No1

WOI HABITAT USE OF OYSTERCATCHERS 921

1995). Therefore, speciesmay selectnesting sites sisted of rocky and sandy beaches. On Flinders that allow efficient accessto territories with suit- Island, sandy beachespredominated on the east- able feeding substratesand prey. em and southwest coast, while rocky beaches Jehl (1985) noted that American Black Oys- predominated on the north, and northwestcoasts; tercatchers(Haematopus bachmani) and Amer- offshoreislands were mainly rocky (Edgecombe ican Oystercatchers(Haematopuspalliatus; a pied 1986). species) nesting in sympatry appeared cryptic The general habitat of each nest was classified when nesting at dark rocky shoresand light sandy as all rocky, all sandy, or mixed beach (predom- beaches respectively. A pied oystercatcher may inately sand with large boulders interspersed).To be more cryptic than a black oystercatcherat a examine how the two species selected nesting light colored beach because of the visual effect substrates with respect to available substrates, created by the countershaded color pattern the proportionsof the distancesof available rocky, (Thayer 1909, Endler 1978). Incident sunlight sandy, and mixed beachesat , Big illuminating the dorsal side of a uniformly dark (a main study area, Figs. 1, 2) and creates a shadow on the belly making its all small islands were compared to the propor- three dimensional shape readily apparent. By tions of nests found on these same substratesat contrast, the white belly of a pied bird createsan these same locations, for each season. optical illusion that obliterates the effectof shad- ow, making individuals appear cryptic with re- FORAGING TERRITORY USE spect to open habitats. Therefore, it may be the At the Fumeaux Islands, Pied and Sooty Oys- case that oystercatchersselect nesting habitats tercatchersgenerally defended feeding territories that maintain crypsis with respect to color pat- in front of their nest sites. On Big Green Island tern. (Fig. 2) in the 198911990 season we examined The purpose of this study is to examine pat- the characteristics of these feeding territories. terns of habitat use for Pied (Haematopus lon- Using aerial photographswe demarcated nest lo- girostris)and Sooty (H. jidiginosus) Oystercatch- cations for measurement and comparison of the ers (a uniformly dark species)nesting at the Fur- total length, rock length, and mud length ex- neaux Islands, , Australia. We examine tending from the high to the low tide line for: (1) habitat use at three scales:general habitats (e.g., each nest, (2) a random site within 100 m radius island choice), territories (e.g., nesting or feeding) of each nest (i.e., a random site at a territory), and nest sites as it is believed that select and (3) 50 random points along the perimeter of habitats in this respective order (Burger 1985a, the island. The coordinates of random siteswere Klopfer and Ganzhom 1985, Sherry and Holmes chosen from a random numbers table. 1985). We present basic information on repro- ductive biology and discusshow factors like the NEST SITE USE location of preferred prey and potential predators For each species we visually estimated seven may have influenced the habitat selectionfor these physical and vegetative characters at nest and sympatric oystercatchers during the breeding random sites (Burger 1985a). A random site was season. selectedwithin a 100 m radius of each nest. Nest sites and random sites for a specieswere com- STUDY AREA AND METHODS pared to examine how habitats were selected. Nest site characterswere compared to examine STUDY AREA AND GENERAL HABITAT USE speciesdifferences in habitat use. Random sites Habitat use data were collected in the austral for each species were compared to determine springandsummerof1988/1989and 1989/1990, whether habitat characteristicswere generally dif- at the Fumeaux Island Group, the , ferent at locations where Pied and Sooty Oys- Tasmania, Australia (4O”OO’S, 148”OO’E). Data tercatcherswere breeding. were collected for Pied and Sooty Oystercatcher In the 1988/1989 seasonon Big Green Island nests found on Flinders Island, the large island the tone of all nest sites and random locations of the group (137,430 ha, Edgecombe 1986) and around nest sites was measured to examine for eight small offshoreislands (all less than 200 whether the two speciesmatched their color pat- ha; Fig. 1). tern to the color of the nesting area. All substrates The shorelines of the Fumeaux Islands con- and speciesof vegetation within a 1 m and 5 m 922 BROOK LAURO ANDERICA NOL

ISLAND

6 tFERGIJSON5 i \ BIG GREEN ISLAND p

FIGURE 1. A map of the Fumeaux Islands,Australia. The eight small islandssurveyed were: Big Green Island,an unnamedisland (1) Chalky(2) Mile (3) Isabella(4) Little Chalky(5) Kangaroo(6) Little Anderson (7), Tin Kettle (8) and Little Green (9). radius of nest and random sites were ranked on covered by substratesand speciesof vegetation a tonal scale of l-10 (l-white and lo-black). A by their color ranking and then summing them. color index at a site was calculated for 1 and 5 Several habitat characterswere used to assess m circles by multiplying the proportion of area nest site visibility. The percent of each substrate HABITAT USE OF OYSTERCATCHERS 923

rocky intertidal l Sooty nest

0 Pied nest

BIG GREEN ISLAND

-N 200 m

FIGURE 2. A map of nest sitesfor Sootyand Pied Oystercatcherson Big Green Island in the 198911990 season. and the percent of each speciesof vegetation were than the Mann Whitney U-tests which were used measured within a 1 and 5 m radius of the sites. for all other comparisons. The categories of substrates were rock, sand, We examined whether the distance of nests to wrack (dead sea weed washed onto the shore) those of potential competitors, both conspecific and other. The angle of view (AOV) from nest and congeneric,influenced the choice of nest site sites was estimated as the degreesof a circle vis- visibility. This was done in the 1989/1990 field ible to a bird sitting on a nest. Nest cover was season by examining the correlations between estimated as the percentage of vegetation cov- angle of view and nearest neighbor distancesfor ering a 0.25 m radius of a circle at a height of nests on Big Green Island (Burger 1977). 50 cm above the ground at the site where eggs Finally, given the potentially high correlation were placed or where a random point was se- between habitat variables we performed a prin- lected. Vegetation distance (Veg. Dst.) was the cipal components analysis using the variables distance to the nearest piece of live vegetation percent substrate and vegetation at 1 and 5 m, from the nest or random site. Vegetation height vegetation distance, vegetation over nest, and (Veg. Ht.) was the height of the nearest species angle of view using the 1988/1989 data to de- of live vegetation. termine if there were: (1) speciesdifferences in Since nestswere either under vegetation or out some combination of these data, and (2) to de- in the open the speciesvalues for percent cover termine whether successfulor unsuccessfulnests and vegetation height were extremes over the differed in their scoreson principal components range of characteristicswith distributions con- 1 or principal components 2 (i.e., Schieck and taining many zero values. Therefore, chi-square Hannon 1993). For the 1990 data we also added analyseswere used for testsof significancerather the variable ‘locale visibility’, which was a mea- 924 BROOK LAURO AND ERICA NOL

sure of the long distance visibility to determine Oystercatchersnesting more commonly at sand- if this variable added significantly to the varia- ier locations and Sooty Oystercatchers nesting tion explained by the principal components. For more commonly at rockier locations. locale visibility, open beaches, intermediate On Flinders Island, Pied Oystercatchersnested beaches, and coves were coded as 1, 2 and 3 at sandy beachesonly and not at rocky or mixed respectively. beaches although these habitats were available (Table 1). On small islands the proportions of REPRODUCTIVE BIOLOGY their nests on mixed beach and rock beach (all sand beaches were rare) were greater and less Information on the reproductive biology for Pied (respectively)than that which would be expected and Sooty Oystercatcher pairs was recorded on given the habitats available (1988/1989: x2 = Big Green Island in both seasons.Clutch sizes 29.00, df = 1, P < 0.001; Table 1). In addition, were recorded for first nests and renests. A pair the number of breeding pairs per km for Pied that hatched at least one egg (including renests) Oystercatcherson Flinders Island was low at 0.5 was considered as successfullyhatching. A pair pairs per km compared to that on Big Green that fledgedat least one chick was categorizedas Island where the number of breeding pairs per successfullyfledging. Where possible we record- km at sandy, mixed beach, rock beach and over- ed reasonsfor egg and chick loss including tidal all was: 4.8,8.8, 1.4 and 3.1 per km (respectively, flooding, severe weather and predation. Table 1, Fig. 2). Sooty Oystercatchers did not nest at sandy POTENTIAL PREDATORS beachesbut did nest at rock and mixed beaches Throughout the first field season(25 September (Table 1). The proportion of nestson rock beach 1988-3 1 January 1989), censuseswere conduct- and mixed beach was not significantly different ed on Flinders Island and Big Green Island to than that which would be expected given the document the presenceof potential predators of habitats available (1989/1990: x2 = 0.44, df = Pied and Sooty Oystercatchersduring daylight 1, P > 0.05; Table 1). In addition, the number hours. We recorded every half-hour over a pe- of breeding pairs per km at rock beach, mixed riod of several hours, the number of visible, po- beach and overall was 4.6, 7.5 and 4.9 pairs re- tential predatorson nesting beaches.On Flinders spectively (Fig. 2). Island, observationswere conductedat four sandy The availability of substratesin the intertidal beaches:Blue Rocks, the Bluff, Ferguson’s Jetty region (Figs. 2,3) influenced the choice of nesting and the Patriarch Inlet (Fig. 1). On Big Green locations by Pied Oystercatchers at Big Green Island observationswere conductedat two mixed Island. The frequency of Pied Oystercatchernests (rocky and sandy) beaches on the eastern shore with respectto beach/intertidal habitats was sig- and at three rocky beacheson the western shore nificantly different than that which would be ex- (Figs. 1, 2). In addition to censusing potential pected based on the habitats that were available predators at half hour intervals the frequency of (Fig. 3; x2 = 17.58, df = 3, P < 0.005). The all potential predators that were seen less regu- proportion of Pied Oystercatchernests on mixed larly (i.e., raptors and ravens) was recorded beach/mixed intertidal and rock beach/rock in- throughout the observation periods. tertidal was greater and in near equal proportion (respectively) to that which was available (Fig. RESULTS 3). Oystercatchersdid not nest at locations where rock was the only substratepresent on the beach GENERAL HABITAT CHOICE and in the intertidal region (Fig. 3). Only Pied Oystercatchersnested on Flinders Is- At Big Green Island, Sooty Oystercatchers land, the largest island of the Fumeaux group, nested at most combinations of beach/intertidal whereas both speciesnested on the smaller sur- habitats and the frequency of nests with respect rounding islands (Table 1, Fig. 1). The frequency to beach/intertidal habitat was not significantly of nests for the two specieson rocky, sandy, and different than that which would be expectedbased mixed habitat (primarily sand with large boul- on the habitats that were available (Figs. 2, 3; x2 ders interspersed) was significantly different = 3.40, df = 3, P > 0.05). They did not nest at (1988/1989:x2= 3949,df= 2, P < 0.001; 1989/ two small sand beach/mud intertidal locations. 1990: x2 = 29.89, df = 2, P < 0.001) with Pied However, since these locations made up a small HABITAT USE OF OYSTERCATCHERS 925

TABLE 1. A comparisonof the proportionof availablesubstrates to the proportionof nestson eachsubstrate by year.a

TOW Rock (%) Sand(%) Mixed (%) FlindersIsland All beaches(km) 225.6 67.7 (30) 110.5 (70) 0.0 (0) Beachessurveyed (km) 40.9 10.2 (25) 30.1(75) 0.0 (0) Pied nests (1988/89) 16 0 (0) 16 (100) cl (0) Pied nests (1989190) 18 0 (0) 16 (100) 0 (0) Big Green Island Beaches(km) 7.1 5.3 (74) 0.2 (3) 1.6 (23) Pied nests (1988/89) 14 4 (29) 1 (7) 8 (64) Pied nests (1989/90) 8 (36) 1 (5) 3 (59) Sooty nests (1988/89) ;: 19 (68) 0 (0) 9 (32) Sooty nests (1989/90) 36 24 (67) 0 (0) 12 (33) All small islands’ Beaches(km) 30.8 25.0 (8 1) 0.0 (O)b 5.8 (19) Pied nests (1988189) 14 4 (31) 0 (0) 9 (69) Sooty nests (1988/89) 44 34 (77) 0 (0) 10 (23) a Eight small islandsincluding Big Green Island (BGI) weresurveyed in 19880989, only Big Green Island and FlindersIslands were surveyed in 1989/1990. b On sure s of small islandsit wasobserved that mostbeaches labelled sand on mapswere mixed beaches.Therefore, for this com!xaisonbeaches labelled saridy on mapswere assumed to be mixed. Seemethods for further details. proportion of the island’s coastline it was not differences were found for rock length (Z = 0.23, possible to statistically assess whether they P=0.8124)ormudlength(Z=0.63,P=0.5270). avoided this habitat at this location. Intertidal mudflats extended out further at low tide than did rocky intertidal habitats (Figs. 2, FORAGING TERRITORY USE 4). Therefore, since Pied Oystercatchers chose At Big Green Island, Pied Oystercatchersselect- muddier sites,the total length of transectsin front ed locations that had significantly longer ex- of their nests was significantly longer than ran- pansesof mud for foraging territories in front of dom locations around the perimeter of the island nest sites compared to random sites along the (Fig. 4; Z = 2.00, P = 0.0456), while no differ- perimeter of the island (Figs. 2, 4; Mann-Whit- ences were found for the same comparison for ney U-test: 2 = 2.53, P = 0.0114). When nest Sooty Oystercatchers(Z = 1.28, P = 0.2018). In sites and random sites at territories were com- addition, the length of substrate at Pied Oyster- pared no differences were found for length of catcher nest sites and at random sites at terri- mud (Z = 1.67, P = 0.0945) or for length of rock tories was greater than that for Sooty Oyster- (Z = - 1.11, P = 0.2769). Pied Oystercatchers, catchers (nest: Z = 2.43, P = 0.0156; random: compared to Sooty Oystercatchers,selected nest Z = 2.38, P = 0.0175; Fig. 4). siteswith more mud (Z = 3.43, P = 0.0006) and less rock in the intertidal zone (Z = -2.32, P = NEST SITE USE 0.0206); random site comparisons for species Habitat color. The rocky and sandy habitats at were not significantly different for rock (Z = the Fumeaux Islands were light colored and -0.28, P = 0.7821) but were different for mud ranked as a light grey on a scale of 10 (l-white, (Z = 2.45, P = 0.0144). lo-black; Table 2). Therefore, neither Pied nor Sooty Oystercatchers nested around the pe- Sooty Oystercatchers had the opportunity to rimeter of Big Green Island irrespective of the match plumage color to nest site color. Neither length of rock or mud substratein the intertidal species selected nest sites that were darker or region. When nest sites and random sites around lighter than random locations and no speciesdif- the perimeter of the island were compared, no ferenceswith respectto the color of nesting sub- significant differenceswere found for rock length strate were found (Table 2). (Fig. 4; Z = 0.69, P = 0.4885) or mud length (Z Visibility. Both speciesselected nest sites that = 0.08, P = 0.9302). When nest sitesand random were more open than surrounding habitat. Nest sites at territories were compared, no significant sites had more substrate (rock, sand, or wrack) 926 BROOK LAURO AND ERICA NOL

60 n SOOTY q PIED q AVAILABLE 1 50

40 1

R-R R-M M-M S-M FIGURE 3. A comparison of the percentageof nests for Sooty Oystercatchersand for Pied Oystercatchersat different beach/intertidal habitats (rock/rock: R-R, rock/mud: R-M; mixed/mud: M-M; sand/mud: S-M) relative to the percentagethat was available on Big Green Island in 198911990.Since mixed beacheswith rocky intertidal areaswere rare they were groupedinto the categoryrock beach/rock intertidal. All other combinations of beach/ intertidal substratesnot mentioned on the table (e.g., rock beach/mud intertidal) did not occurat Big Green Island. than vegetation as compared to random sites. spite the inclusion of an additional variable in Differences were significant for Pied Oyster- 1990. In both years, positive values of PC1 ap- catchers within a 5 m radius of nests and for peared to represent greater proportions of sub- Sooty Oystercatcherswithin a 1 and 5 m radius strate, less vegetation around nest, little to no of nests (Table 2). vegetation over the nest and a low value for angle Within open areas, Sooty Oystercatchersusu- of view whereas negative values represent more ally placed nests next to and under a clump of enclosed sites with more vegetation near nests, vegetation while Pied Oystercatchers usually lesssubstrate, and more vegetation over the nest. placed nestsin the open. Pied Oystercatchernest PC2 had a particularly high negative loading on sites had a significantly higher angle of view and the variable of locale visibility, which suggests lower percentcover than random locations. Sooty that PC2 also measured the tendency to nest in Oystercatchersselected sites where distances to coves (locale visibility = 3), without the long the closestpiece of vegetation were significantly distance visibility of beaches(locale visibility = closer than random locations (Table 2). In ad- 1). dition, Sooty Oystercatcher nest sites had a sig- When we compared the averagevalues of PC1 nificantly lower angle of view, higher percent nest and PC2 for the two species,they were, as ex- cover, and shorter vegetation distance than Pied pected, significantly different in both years (Ta- Oystercatcher nests while the comparisons for ble 4). Sooty Oystercatchers had significantly the random charactersfor the two specieswere lower values for PC1 and PC2 than did Pied not significantly different (Table 2). Oystercatchers.This confirms the univariate re- In the 1989 and 1990 seasons,PC1 and PC2 sults that Pied Oystercatchersnest in the open, togetherexplained 70% and 66% of the variation with less vegetation or no vegetation over their in the data respectively (Table 3). The loadings nests, while Sooty Oystercatchershave both re- were similar in direction in 1989 and 1990, de- duced local and long distance visibility. HABITAT USE OF OYSTERCATCHERS 921

120

100

z 8o X ; 60 Z w Ll 40

I-R S-N S-R P-N P-R FIGURE 4. A comparisonon Big Green Island in 1989/1990 of the averagelength (m) of rock and mud betweenthe hightide line and the low tide line at: randomsites along the islandperimeter (I-R, n = 50); Sooty Oystercatchernest sites (S-N, n = 33) and randomsites at nestingterritories (S-R, n = 33); Pied Oystercatcher nestsites (P-N, IZ= 17) and randomsites at nestingterritories (P-R, n = 17). Errorbrackets are f one standard error.

Finally, when conspecificswere near, both spe- Both speciesplaced nests on a soft substrate cies chose sites with lower angles of view, per- that cushioned eggs.Sooty Oystercatcherscom- haps reducing the potential for intraspecific con- monly placed nests on wrack apparently because flict. Nearest conspecificneighbor distancesand it was the primary soft substrateavailable at rocky angle of view were positively and significantly habitat. Therefore, Sooty Oystercatchersselected correlated for Pied (r = 0.69, P = 0.0023) and sites with more wrack than random sites (Table Sooty Oystercatchers(r = 0.73, P = 0.0001) while 2). Twenty four (63%) Sooty nests were placed weaker positive and nonsignificant correlations on a base of wrack, eight (20%) were on dead were found for congenericneighbor distancesfor vegetation, three (9%) were on rock, two were Sooty (r = 0.33, P = 0.1301) and Pied (r = 0.46; (5%) on sand and one was (3%) on pigface (Dis- P = 0.06 19) Oystercatchers. phyma australe). Twelve (42%) Pied Oyster- catcher nests were placed on a base of sand, ten NESTING SUBSTRATES (33%) were on wrack, four (13%) on dead veg- Species differences with respect to rocky and etation, one (4%) was on rock, one (4%) was on sandy substrates were found at nest sites and shells and one (4%) was on pigface. The fre- random sites and were related to their general quency of Sooty Oystercatcher nests on a base habitat preferences. Sooty Oystercatchers,who of wrack and sand was significantly higher and nestedmore commonly at rocky shores,had more lower respectively than for Pied Oystercatchers rock within a 5 m radius of nests compared to (wrack: x2 = 5.953, df = 1, P < 0.025; sand: x2 random sites (Table 2). By contrast, Pied Oys- = 15.30, df = 1, P < 0.005). tercatchers,who nestedmore commonly at sandy beaches,had significantly greater percentagesof REPRODUCTIVE BIOLOGY sand within a 1 and 5 m radius of nests as com- Clutch size.At Big Green Island the modal clutch pared to random sites (Table 2). size for first nests for each specieswas two eggs. 928 BROOK LAURO ANDERICA NOL

TABLE 2. Comparisonsof habitatcharacteristics of Sooty(n = 38) and Pied Oystercatcher(n = 29) nestsand random siteswithin 100 m of nests,in 1988/1989.

Comparisons’ SootyOystercatcher PiedOystercatcher Sooty-Pied Nest Random Nest Random Nest-Random Ran- R (SE.) x (SE.) R (S.E.) x (S.E.) sooty Pied Nest dom 1 m Circle: Color Indexb 3.1 (0.20) 3.2 (0.25) 2.9 (0.26) 2.8 (0.44) ns ns ns ns % Vegetation 32.6 (3.13) 55.7 (6.83) 24.4 (4.78) 48.1 (8.06) * ns ns ns % Substrate 67.4 (3.11) 44.3 (6.81) 75.6 (4.78) 51.9 (8.06) * ns ns ts % Rock 32.3 (4.50) 27.8 (6.33) 9.8 (3.91) 9.3 (5.11) ns ns *** % Sand 9.1 (2.94) 9.9 (3.51) 38.1 (5.88) 25.1 (7.14) ns * *** ns % Wrack 18.8 (3.63) 1.5 (0.93) 21.6 (5.16) 12.6 (5.36) *** ns ns ns % Other 6.7 (1.60) 5.1 (2.23) 6.1 (2.29) 4.9 (2.07) ns ** ns ns 5 m Circle: Color Indexb 2.8 (0.13) 3.1 (0.23) 2.8 (0.23) 2.5 (0.82) ns ns ns * % Vegetation 35.8 (2.73) 59.7 (6.63) 27.7 (4.91) 57.0 (7.30) ** ** ns ns % Substrate 64.2 (2.73) 40.3 (6.63) 72.3 (5.06) 42.9 (7.25) ** ** ns ns % Rock 38.2 (3.85) 27.4 (5.95) 15.6 (5.01) 6.7 (3.02) ns *** ** % Sand 9.7 (2.64) 5.9 (1.96) 41.8 (6.47) 26.2 (6.72) :s *** *** * % Wrack 8.8 (1.73) 4.5 (2.77) 11.6 (3.24) 8.6 (3.57) *** ns ns ns % Other 7.4 (2.57) 2.5 (0.97) 3.3 (1.83) 1.4(1.38) ** ** ns ns AOV (“) 87.0 (11.47) 144.9(26.00) 219.2 (13.38) 140.0(25.79) ns * *** ns % Nest Covert 25.5 (6.22) 20.8 (6.54) 0.1 (0.07) 13.0(6.16) ns * *** ns Veg. Dst. (cmp 25.9 (15.24) 103.9(33.19) 140.6(42.0) 99.1 (36.7) ns ns Veg. Ht. (cm) 84.6 (14.92) 46.5 (12.13) 45.4 (11.8) 62.4(11.4) is ns is ns

pMann-Whitney, NS = I’ > 0.05;* = P < 0.05;** = P < 0.01;*** = P < 0.001. b Data werefor SootyOystercatcher nests (n = 30) and PiedOystercatchers nests (n = 18)located on BigGreen Island in thesecond field season. r Chi-squarecontingency table analyses were conducted on categories.

For Pied Oystercatchers,the mean clutch size for the mean clutch size for first nestsin the first and first nests in the first and second field seasons secondfield seasonswas 1.7 (ltO.09, n = 26) and was 1.7 (k0.21, n = 10) and 1.9 (kO.06, IZ = 15) 2.0 (kO.03, n = 30; respectively). The range for while the range in each year was l-3 eggsper clutch size in both years was l-2 eggsper clutch; clutch. No renests were documented in the first no three egg clutches were observed in either seasonbut in the second seasonthree were doc- seasonand in the second seasononly one clutch umented with clutch sizes of 1, 2 and 2 eggsper had one egg. In each field seasonone renest was clutch. Pied Oystercatchersusually only raised documented and the clutch size in 1988/1989 one chick but in the first seasonone pair fledged was one egg while the clutch size in 1989/1990 two chicks and in the second season two pairs was two eggs. Sooty Oystercatcherswere never each fledged two chicks. observed to raise more than one chick. For Sooty Oystercatchersat Big Green Island Clutch and brood loss. At Big Green Island,

TABLE 3. Loadingsof habitat variableson first and secondprincipal components (PC1 and PC2) in 1988/ 1989 and 1989/1990.

1989 1990 PCAl PCA2 PCAI PCA2 Substratelm 0.44 0.37 0.50 0.06 Vegetation1 m -0.44 -0.37 -0.49 -0.05 SubstrateSm 0.39 -0.57 0.39 -0.35 Vegetation5m -0.39 0.57 -0.35 0.40 Distanceto veg. 0.33 0.19 0.29 0.18 Veg. over nest(O/o) -0.32 -0.10 -0.35 -0.16 Angle of view 0.30 0.18 0.18 0.57 Localevisibility - - -0.01 -0.58 HABITAT USE OF OYSTERCATCHERS 929

TABLE 4. Principal component scoresfor PC1 and PC2 for Pied and Sooty Oystercatchersfor 1989 and 1990.

Year Variable Sooty Pied t P

1989 PC1 -0.36 (0.379) 0.60 (0.300) 2.00 0.052 PC2 -0.36 (0.191) 0.60 (0.295) 2.88 0.006 1990 PC1 -0.37 (0.378) 0.61 (0.301) 2.02 0.048 PC2 -0.44 (0.200) 0.74 (0.337) 3.22 0.002

in the 1988/1989 season,seven of fourteen first no chick. Of the 14 clutches (including renests) clutches (50%) for Pied Oystercatcherswere lost that hatched, 10 (7 1%) fledgedat least one chick, (Table 5). No nest was known to be flooded by while four (29%) did not fledge a chick and were high tides but one nest on the south coast was presumed depredated. lost because high winds eroded the base of the For Sooty Oystercatchers,in the first and sec- nest. One clutch of eggswas destroyed when a ond season, 10 of 28 (36%) and 16 of 36 (44%; local sheep farmer threw gear onto a nest by respectively) first clutcheswere lost (Table 5). In mistake. No apparent reason was found for the the first season,the one pair that renestedhatched loss of the remaining clutches,but they were pre- one chick which did not fledge. In the second sumed to be depredated (see section below on season the one pair that renested hatched no potential predators). Of the seven clutches to chicks. No nestswere known to be damaged due hatch chicks,four (57%) fledgedat least one chick, to tidal flooding or wind damage. In the first and while three (43%) fledged no chick. No apparent second seasons, for the 18 and 20 pairs that reason for chick loss was observed but pre- hatchedat least one chick, 11(6 1%) and six (30%; dation was suspected. respectively) complete broods were lost (includ- In the 1989/1990 season,eight of twenty-two ing renests). We were unable to document any first clutches(36%) for Pied Oystercatcherswere direct causesfor eggor chick lossesbut predation lost (Table 5). The causesfor clutch loss were as was suspected. follows: two were trampled by sheep, two were Success.There were no overall differences in flooded, one was abandoned due to interspecific hatching or fledging successbetween speciesin competition and three were unknown but pre- either season or for the two seasonscombined sumed depredated.The Pied nest was abandoned (Table 5). Pied Oystercatcher hatching or fledg- just before hatching becauseof a continual attack ing successwith respect to habitat type was not on adults by a pair of Sooty Oystercatchersnest- significantly different in either field season(Table ing twenty meters away. For the three pairs that 5). Sooty Oystercatcherhatching successwith re- were observed to renest, one clutch was hooded, spect to habitat type was also not significantly one was lost for unknown reasonsand presumed different in either field season (Table 5). How- depredated,while one clutch hatched but fledged ever, in the second season,Sooty Oystercatchers

TABLE 5. Pied and Sooty Oystercatcherhatching (H) and fledging (F) successat Big Green Island. The total number of nests by species,year and habitat is followed by the number (percent) hatching and fledging.

Mixed (sandand rock) Rock Sand OVK%ll n H (%) F(s) n H (%) F(s) n H (%) F(s) n H (%) F (%) 1988/1989a sooty 9 5 (56) 3 (33) 19 13 (68) 4 (21) - - - 28 18 (64) 7 (25) Pied 8 4 (50) 2 (25) 5 3 (60) 2 (40) 1 O(0) O(0) 14 7 (50) 4 (29) 198911990’ sooty 12 4 (33) 1 (8)b 24 16 (67) 13 (54)b - - - 36 20 (56) 14 (39) Pied 14 8 (57) 5 (36) 7 6 (86) 5 (71) 1 0 (0) cl(0) 22 14 (64) 10 (45)

a When data werecombined for the two seasonsno differencesoccurred between species for hatching(df = I, x2 = 0.10, P > 0.05) 01 fledging(df = 1, ,x2= 0.37, P > 0.05) success. b Slgnilicantdifferences occurred between mixed and rock habitatsfor SootyOystercatchers in 1989/1990 (df = I, x2 = 7.07, P < 0.01). 930 BROOK LAURO AND ERICA NOL

TABLE 6. Comparison of PC1 and PC2 scores(x, SE) for hatched and unsuccessfulnests of Sooty and Pied Oystercatchersby year. No comparisonswere significantly different.

Species Variable Hatched Did not hatch

198811989 sooty PC1 -0.24 (0.745) -0.08 (0.525) PC2 -0.18 (0.326) -0.13 (0.388) 11 12 Pied kl 0.10 (0.481) 0.68 (0.859) PC2 0.37 (0.458) 0.09 (0.262) n 8 4 1989/1990 sooty PC1 -0.21 (0.379) -1.81 (1.637) PC2 -0.50 (0.218) 0.10 (0.301) 27 Pied kl 0.96 (0.335) i.05 (0.533) PC2 0.63 (0.505) 0.9 1 (0.393) n 11 7

had lower fledging successon mixed habitat than (Table 7). There were no differences in the fre- rocky habitat (Table 5); no differenceswere found quency of Silver Gulls between the mixed and in the first season(Table 5). rocky beachesof Big Green Island (Table 7). Pa- There were no significant differences in the cific Gulls were seenmore frequently on the rocky PC1 and PC2 scores for nests that ultimately shores of Big Green Island than on the sandy hatched and those that did not (Table 6). Thus, shoresof Flinders Island or on the mixed beaches the degree to which the nest was visible (prin- of Big Green Island (Table 7). When the fre- cipally) did not appear to explain differential re- quency data for Silver Gulls on rocky and mixed productive successfor either species. habitat on Big Green Island were combined and compared to the frequency of Silver Gulls on POTENTIAL PREDATORS Flinders Island, there were more Silver Gulls on The average number of half-hour censusesper Flinders Island than on Big Green Island (Mann- day on the sandy beachesof Flinders Island was Whitney U-test: Z = 3.003, P = 0.0027; Table 5.1 (kO.68, n = 19). On Big Green Island the 7). There were no differencesin the frequency of average number of censusesper day on mixed Pacific Gulls on rocky and mixed habitat (com- beachesand rocky beacheswas 4.9 (kO.36, n = bined) on Big Green Island compared to the fre- 26) and 5.1 (kO.52, n = 25) respectively. The quency of Pacific Gulls on Flinders Island (2 = only potential predators observed during these -0.245, P = 0.8062; Table 7). When the fre- daytime censuseswere birds; no mammals or quencies for all potential avian predators were reptiles were recorded. Six species of potential combined for observation periods it was found avian predatorswere seenincluding: Pacific Gulls that there were more individuals found on Flin- (Larus pacificus), Silver Gulls @ants novuehol- ders Island than on Big Green Island (Z = 2.356, landiae), Ravens (Corvus tasmanicus), Harriers P = 0.0056; Table 7). This was due primarily to (Circus aeruginosus),Brown Falcons (Falco be- the greater numbers of Silver Gulls found on the rigora), and Sea Eagles (Haliaeetus leucogaster). beaches of Flinders Island as compared to Big The first three species mentioned would more Green Island. likely be predators of eggsand chicks while the We did not statistically compare the frequen- three remaining species would more likely be cies of Ravens, Brown Falcons, Sea Eagles, and predatorsofjuveniles and possibly adults (Pizzey Harriers occurring either between habitats or be- 1980, Lane 1987). tween locations since individuals for these spe- Both speciesof gull were observed commonly cies were all observed rarely. None of these spe- during behavioral observations while the re- cies were seen more than eight times during ob- maining bird specieswere seen rarely (Table 7). servation periods. Silver Gulls were observed more frequently on A Pacific Gull on one occasion was observed sandy beacheson Flinders Island than on rocky to go after Sooty Oystercatcher eggs while on beachesor on mixed beachesat Big Green Island another occasion a Pacific Gull was observed to HABITAT USE OF OYSTERCATCHERS 931

TABLE 7. A comparison of the mean (standard error) number of potential predators observed in 1988/1989 at Flinders Island and Big Green Island for different beach substrate.

Flinders Island Big Green Island Sand (n = I9 days) Mixed (n = 26 days) Rock (n = 25 days) Species R SE R SE + SE

White Breasted Sea Eagle 0.02 0.018 0.00 0.000 0.00 0.000 Swamp Harrier 0.01 0.009 0.00 0.000 0.01 0.000 Brown Falcon 0.00 0.000 0.01 0.005 0.01 0.010 Silver Gull 5.92 1.354.b 2.15 0.472’ 1.95 0.589b Pacific Gull 1.45 0.327c 0.81 0.201d 2.24 0.349.d Raven 0.01 0.009 0.02 0.019 0.00 0.000 Total 7.41 1.546 2.99 0.603 4.21 0.898

Significance values for Mann Whitney tests: a, Z = 2.54, P < 0.01; b, Z = 2.76, P < 0.01; c, Z = -1.96, P < 0.05; d, Z = 4.19, P < 0.001. go after Sooty Oystercatcher chicks. In both sit- veyed one large island in this region and a survey uations, attacks were unsuccessful. One Sooty on other large islands in the BassStrait (e.g., Cape Oystercatcher egg, which was probably aban- Barren and King) would be valuable. doned, was found to be eaten by a rat. At the Fumeaux Islands, Sooty Oystercatchers nested more commonly at rocky shores while DISCUSSION Pied Oystercatchersnested more commonly at Pied and Sooty Oystercatchersexhibited differ- sandy shores. Black and pied speciesof oyster- encesin habitat use at the levels of general hab- catchersmay selectnest sites at rocky and sandy itats (islands,substrates), territories and nest sites, habitats (respectively) since these habitats are ecologically segregatingand presumably facili- commonly dark and light in color (respectively) tating their coexistenceat the Fumeaux Islands, hence aiding in maintenance of nest site crypsis Australia. Sooty Oystercatchersnested on small relative to color pattern (Jehl 1985). However, islands only, and although nesting and feeding at the Fumeaux Islands, Sooty Oystercatchers territories appeared to be available on Flinders nested at rocky habitat even though they were Island (a large island), they did not nest there. light and not significantly different in color from Pied Oystercatchersnested on all sized islands. sand beaches. In addition, the number of breeding pairs per km One factor at the Fumeaux Islands that was was higher for both speciesat Big Green Island, related to rock versus sand nesting was island a small island, than for Pied Oystercatchersat choice. Sooty Oystercatchersnested only at small Flinders Island. islands that were mainly rocky while Pied Oys- Coastal birds commonly nest at higher den- tercatchersnested on Flinders Island, a large is- sities on small islands as compared to larger bod- land which was mainly sandy, as well as small ies of land (e.g., large islands or mainland sites) islands. Another factor involved was the location and two important factors that may contribute of nesting areasrelative to preferred feeding sub- to this pattern are reduced predation risk and strates.Pied Oystercatchernesting areaswere re- lower human disturbance at smaller islands as stricted to locations that had intertidal mudflat compared to larger islands (Buckley and Buckley for feeding. At intertidal mudflats, Pied Oyster- 1980, Williamson 1981, Blonde1 1985, Nilsson catchers fed primarily on soft bodied prey and et al. 1985, George 1987, Erwin et al. 1995). they were not observed to capture prey off rocks Censusesconducted for this study did show that at any location (Lauro and No1 1995). The lo- gulls, known predators of oystercatchereggs and cation of suitable foraging substratesdid not ap- chicks (Harris 1967, Heppleston 1972, Hartwick pear to limit where Sooty Oystercatchersnested; 1974, Hockey 1983, Safiiel 1985, No1 1989), were they selected sites that had mud or rock at ad- more common at the shoreline of Flinders than jacent intertidal areas. This pattern was related at the shoreline of Big Green Island. In addition, to the fact that Sooty Oystercatchers captured human disturbance would have been higher on high proportions of hard-shelled prey at inter- Flinders as it was inhabited by people while the tidal mudflats and at rocky shores (Laura and small islands were not. However, we only sur- No1 1995). 932 BROOK LAURO AND ERICA NOL

An important difference in nest site use for the though predation on adults presumably would two species was that Sooty Oystercatchers se- be low, our data on reproductive successsuggest lected lower visibility sites compared to Pied predation on chicks and eggswas probably oc- Oystercatchers.There are several possibleexpla- curring. Thus, at the Furneaux Islands, where the nations for this pattern which may not be mu- rocky and sandy beacheswere generally light in tually exclusive. Sooty Oystercatchersmay have color, Sooty Oystercatchersmay compensate for nested in lower visibility locations becausethey their uniformly dark and presumably non-cryp- were seeking sites which provided shade. Dark tic color pattern by selecting low visibility nest and light colored birds experience different heat sites. In contrast, Pied Oystercatchers with a loads which may affect behavior (Walsberg et al. countershadedand presumably more cryptic col- 1978). Therefore, the uniform black plumage of or pattern nested in higher visibility nest sites. Sooty Oystercatchers may have caused higher This pattern of nest site use would be especially head loads than the countershaded plumage of important in avoiding predators that hunt vi- Pied Oystercatchers.Yet, both speciesare dark sually during the day, such as gulls. Silver Gulls from above and may therefore experiencesimilar and Pacific Gulls rested, fed and nested within dorsal heat loads. In addition, thermoregulatory and about oystercatcherterritories. Pacific Gulls stress for the two specieswas rarely witnessed were observed to attack oystercatchereggs and during two field seasonsof behavioral observa- chicks. tions. The Bass Strait Islands experience a tem- Other evidence suggestedthat predation was perate climate with high winds and daily sea influencing patterns of nest site use for both spe- breezes (Edgecombe 1986) that cool incubating ties. Both speciesselected areas about nest sites birds. Wind negates heat load differences asso- that were more open and that were closer to the ciated with different color patterns of birds waters edge than randomly selectedsites (Lauro (Walsberg et al. 1978). and No1 1993, Lauro 1994). Open sites around Sooty Oystercatchersmay have nested in low- nests provide adults with the opportunity to de- er visibility sites to avoid interspecific conflict tect predators before eggsor chicks can be dis- with Pied Oystercatchers.If this was the case, it covered (Burger 1977, Burger and Gochfeld would be expectedthat when Sooty Oystercatch- 1988). Hartwick (1974) showed that American ers nested close to Pied Oystercatchersthe vis- Black Oystercatchersnested at open sites, close ibility about their nest siteswould be lower than to the waters edge, even though nests were com- when they nested at greater distancesfrom Pied monly flooded at these locations. He suggested Oystercatchers.For certain speciesof gulls, the that nesting pairs selectedthese sites because at closerthe congenericor conspecificneighbor, the higher elevated sites, in denser vegetation at dis- lower the visibility at the nest site and when tances further from the waters edge, eggs and vegetation about nests was experimentally re- chicks were more susceptible to predation by moved interspecific and intraspecific interaction nesting gulls. increased(Burger 1977). However, for this study Interspecific competition (past or present day) there was no significant correlation between may have had subtle influences on specieschoice nearest neighbor distance and angle of view for of nesting habitats. For example, the fact that congenericsalthough there was a significant pos- Sooty Oystercatcherswere not observed to nest itive correlation for conspecifics. on sand beaches and Pied Oystercatcherswere A more tenable explanation for speciesdiffer- not observedto nest at rock intertidal/rock beach ences in choice of nest site visibility is that spe- suggestscompetitive exclusion. However, pres- ties selectedvegetative characteristicsthat min- ently it seems unlikely that Pied Oystercatchers imized predation risk with respect to color pat- excluded Sooty Oystercatchersfrom any habitat tern. Numerous studies have shown that breed- since they were smaller and appearedto be com- ing birds select vegetative characteristicsat nest petitively inferior. At mixed beaches,where both sites, such as visibility, to reduce potential risks speciesnested side by side (Fig. 2), Sooty Oys- of predation (Burger 1977, 1985a, 1985b; Yah- tercatchersregularly chasedPied Oystercatchers ner and Cypher 1987; Burger and Gochfeld 1988; from desired foraging areas, and in one case a Watts 1990; Holway 199 1; Pingjuin and Martin Sooty Oystercatcher pair excluded a Pied Oys- 1991;Martin 1988, 1992;Kelly 1993;Lima 1993; tercatcherpair from nesting; Pied Oystercatchers Norment 1993; Schieck and Hannon 1993). Al- rarely chasedSooty Oystercatchersat these sites. HABITAT USE OF OYSTERCATCHERS 933

It also seemsunlikely that Sooty Oystercatchers EDGECOMBE,J. M. 1986. FlindersIsland and Eastern excluded Pied Oystercatchersfrom rock beach/ BassStrait. J. M. Edgecombe,Sydney, Australia. ENDLER,J. A. 1978. A predator’sview of animalcolor rock intertidal locations for nesting. On Flinders patterns.Evol. Bio. 11:319-364. Island, where no Sooty Oystercatchers nested, ERWIN.M.. J. HA~FTELD.AND T. WILMERS. 1995. The Pied Oystercatchersdid not nest at rock beach/ value and vulnerabilityof small estuarineislands rock intertidal areasalthough thesehabitats were for conserving metapopulations of breeding wa- available. Removal experiments would help to terbirds. Biol. Cons. 7 1:187-l 9 1. FRANK,P. W. 1982. Effectsof winter feeding on lim- clarify the question of whether present-day in- pets by Black Oystercatchers,Haematopus bach- terspecific competition influences species pat- mani. Ecology 63:1352-1362. terns of habitat use. GEORGE,T. L. 1987. Great land bird densities on island vs. mainland: relation to nest predation lev- el. Ecology 68:1393-1400. ACKNOWLEDGMENTS Goss CUSTARD,J., R. CALDOW,AND R. CLARKE. 1992. Correlatesof the density of foragingoystercatchers Specialthanks go to Derek Smith, Garth Smith, Bobby Haematopus ostralegusat different population Gardner, Duke Goss, and Margaret Goss for their as- sizes. J. Anim. Ecol. 61:159-173. sistancein the field at the Fumeaux Islands. We are HARRIS,M. P. 1967. The biology of oystercatchers grateful to Allan Baker who suggestedthat Flinders Haematopusostralegus on SkokholmIsland, South would be a suitable site for field work and who sug- Wales. Ibis 109:180-193. gestedcontacting Derek Smith. This work wasfunded HARTWICK,E. B. 1974. Breedingecology of the Black bv an NSERC (Canada)grant to E. No1 and Oueen’s Oystercatcher(Haematopus bachmani Audubon). University GraduateFkIiowships and Awards to B. Syesis7183-92. Lauro. We appreciatethe help of Barb Mountney in HARTWICK,E. B. 1976. Foragingstrategy of the Black the preparation of the manuscript. Oystercatcher(Haematopus bachmani Audubon). Can. J. Zool. 54:142-155. LITERATURE CITED HAYMAN,P. J., J. MARCHANT,AND T. PLATER. 1986. Shorebirds-an identification guide. Houghton BAKER,A. J. 1974a. Ecological and behavioral evi- Mifflin, Boston. dence for the systematic status of New Zealand HEPPLESTON,P. B. 1972. The comparative breeding oystercatchers (: Haematopodi- ecologyof oystercatchers(Haematopus ostralegus dae). Lif. Sci. Contr., Roy. Ont. Mus., no. 96. L.) in inland and coastal habitats. J. Anim. Ecol. BAKER,A. J. 1974b. Prey-specific feeding methods 41:23-51. of New Zealand ovstercatchers.Notomis 2 1:219- HOCKEY,P.A.R. 1982. Adaptiveness of nest site se- 233. lection and egg coloration in the African Black BENT,A. 1929. Life historiesof North American OystercatcherHaematopus moquini. Behav. Ecol. shorebirds,Part II. SmithsonianInstitute. United and Sociobio. 11:117-l 23. States National Museum, Bulletin 146. HOCKEY,P.A.R. 1983. The distribution, population BLQNDEL,J. 1985. Habitat selectionin island versus size, movements and conservationof the African mainland birds, v. 477-516. In M. Codv Ied.1. Black Oystercatcher Haematopus moquini. Bio. Habitat selection-inbirds. AcademicPress, New Cons. 25~233-262. York. HOCKEY,P.A.R., AND G. M. BRANCH. 1984. Oyster- BUCKLEY,F. G., AND P. BUCKLEY. 1980. Habitat se- catchersand limpets: impact and implications-a lection and marine birds. p. 69-l 12. In J. Burger, preliminary assessment.Ardea 72: 199-206. B. Olla, and H. E. Winn [eds.],Behavior of marine HOCKEY,P.A.R., AND L. G. UNDERHILL. 1984. Diet -current perspectives in research, Vol- of the African BlackOystercatcher on rocky shores: ume 4-Marine birds. Plenum, New York. spatial, temporal and sex related variation. S. Afr. BURGER,J. 1977. Role of visibility in nestingbehav- J. Zool. 19:1-11. iour of Larus gulls. J. Comp. Physiol. Psych. 9 1: HOLWAY,D. 1991. Nest site selection and the im- 1347-135s. portanceof nest concealmentin the Black-Throat- BURGER,J. 1985a. Habitat selection in temperate ed Blue Warbler. Condor 93:575-581. marsh-nestingbirds, p. 253-28 1. In M. Cody [ed.], HULXHER. J. B. 1985. Growth and abrasion of the Habitat selection in birds. Academic Press, New oystercatcherbill in relation to dietary switches. York. Neth. J. Zool. 35:124-154. BURGER,J. 1985b. Nest site selection by Laughing JEHL,J. R., JR. 1985. Hybridization and evolution Gulls: comparison of tropical colonies (Culebra, of oystercatcherson the Pacific Coast of Baja Cal- Puerto Rico) with temperate colonies (New Jer- ifornia. Omithol. Monogr. 36:484-504. sey). Condor 87:364-373. KELLY,J. 1993. The effect of nest predation on hab- BURGER,J., AND M. GOCHFELD.1988. Nest site se- itat selectionof Dusky Flycatchersin limber pine- lection-by Roseate Terns in two tropical colonies juniper woodland. Condor 95:83-93. on Culebra. Puerto Rico. Condor 90:843-85 1. KLQPFER, P. H., ANDJ. U. GANZHORN.1985. Habitat DARE, P. J., A& A. MERCER. 1973. Foods of the selection:behavioral aspects,p. 435-453. In M. Oystercatcherin Morecambe Bay, Lancashire.Bird Cody [ed.], Habitat selection in birds. Academic Study 20:173-184. Press,New York. 934 BROOK LAURO AND ERICA NOL

LANE, B. A. 1987. Shorebirds of Australia. Nelson Sparrows in the Northwest Territories, Canada. Publishers,Melbourne, Australia. Auk 110:769-777. LAURO.B. 1994. Patterns of habitat use for Pied and P~NGJUIN,L., AND T. E. MARTIN. 1991. Nest site Sooty Oystercatchersat the Fumeaux Islands,Tas- selectionand nestingsuccess of cavity-nestingbirds mania, Australia. Ph.D.diss., Queen’s University, in high elevation forest drainages.Auk 108:405- Kingston, Ontario, Canada. 418. LAIJRO,B., ANDJ. BURGER. 1989. Nest site selection Przzav, G. 1980. A field guide to the birds of Aus- of American Oystercatchers(Haematopus pallia- tralia. Princeton Univ. Press, Princeton, NJ. tus) in salt marshes. Auk 106:185-192. SAFRIEL,U. N. 1985. ‘Diet dimorphism’ within an LAURO,B., ANDE. NOL. 1993. The effect ofprevailing oystercatcher population-adaptive significance wind direction and tidal flooding on the repro- and effects on recent distribution dynamics. Ibis ductive successof Pied Oystercatchers.Emu 93: 127:287-305. 199-202. SCHIECK, J., ANLIS. HANNON. 1993. Clutch predation, LAUROB., ANDE. NOL. 1995. Feeding behavior, prey cover, and the over dispersion of nest of the Wil- selection and bill size of Pied and Sootv Ovster- low Ptarmigan. Ecology 74:743-750. catchersin Australia. Wilson Bull.: in press: SHERRY,T. W., ANDR. T. HOLMES. 1985. Dispersion LIMA, S. L. 1993. Ecoiogical and evolutionary per- patternsand habitat responsesof birds in northern spectiveson escapefrom predatory attack: a sur- hardwood forests. D. 283-309. In M. Codv led.1, vey of North American birds. Wilson Bull. 105: Habitat selection’& birds. Academic Press,New l-47. York. LINDBERG,D. R., K. I. WARHE~T,AND J. A. ESTES. SWENNEN,C., L. BRULTN,P. DUIVEN,M. LEOPOLD,AND 1987. Prey preference and seasonalpredation by E. MAR~~JN. 1983. Differences in bill form of oystercatcherson limpets at San Nicolas Island, the osytercatcherHaematopus ostralegus;a dy- California, U.S.A. Mar. Ecol. Prog. Ser. 39:105- namic adaptation to specific foraging techniques. 113. Neth. J. Sea. Res. 17:57-83. MARTIN,T. E. 1988. Processesorganizing open-nest- THAYER,G. H. 1909. Concealing-colorationin the ing bird assemblages:competition or nest preda- kingdom: an exposition of the laws of dis- tion. Evol. Ecol. 2:37-50. guisethrough color and pattern. Macmillan, New MARTIN, T. E. 1992. Interaction of nest predation York. _ and food limitations in reproductive strategies. YAHNER,R., AND B. CYPHER. 1987. Effects of nest Cum Omith. 9:163-197. location on depredationof artificial arborealnests. NIISSON, S. G., C. BJORKMAN,P. FORSLUND,AND J. J. Wildl. Manag. 5 1:178-l 8 1. HOGLUND. 1985. Egg predation in forest bird WAL~BERG,G. E., G. CWBELL, ANDA. KING. 1978. communities on islands and mainland. Oecologia Animal coat color and radiative heat gain: a re- 66511-515 evaluation. J. Comp. Physio. 126:21l-222. NOL, E. 1989. Food supply and reproductive perfor- WAITS, B. D. 1990. Cover use and predation related mance of the American Oystercatcherin Virginia. mortality in Song and Savannah Sparrows. Auk Condor 91:429-435. 107~775-778. NORMENT,C. 1993. Nest site characteristicsand nest WILLIAMSON,M. 1981. Island populations. Oxford predation in Harris’ Sparrowsand White Crowned Univ. Press, Oxford, U. K.