Multispecies Seabird Feeding Flocks in the Galapagos Islands’

The Condor 100:277-285 0 The Cooper Ornithological Society 1998

MULTISPECIES SEABIRD FEEDING FLOCKS IN THE GALAPAGOS ISLANDS’

KYRA L. MILLS Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697

Abstract. I examined the speciescomposition, frequency, distance offshore, and duration of multispeciesseabird feeding flocks in the GalkpagosIslands, Ecuador.Flocks were comprised of Galapagos Penguins (Spheniscus mendiculus), Flightless Cormorants (Cumpso- haelius [Nannopterum] harrisi), Brown Pelicans (Pelecanus occidentalis), Brown Noddies (Anous stolidus), Blue-footed (Sula nebouxii) and Masked Boobies (&da ductylatm), Mag- nificent Frigatebirds (Fregata magnz$cens),and Audubon Shearwaters(P@nus Zhermini- eri). Pursuit-divers, suchas GalapagosPenguins and FlightlessCormorants, increase the longevityof flocks.Species such as boobies may tendto dissipateflocks. This studyprovides field evidencethat the presenceof pursuit-divingseabirds has a positiveeffect on the duration of feedingflocks and that the mechanismsthat keep prey closeto the surfacenear shoremay differ from thosein the openocean. Key words: foraging behavior, Galipagos Islands, Galdpagos Penguins, multispecies feeding jocks, seabirds, Spheniscusmendiculus.

INTRODUCTION 1978, 1984, Feldman 1986) allow this area to Seabird flocks are temporary associations in support an abundance of marine mammals and which individuals congregateto feed, sometimes seabirds, including pursuit-diving species such in very large numbers. These flocks are partic- as the Flightless Cormorant (Campsohaelius ularly prominent in tropical waters (Duffy 1983, [Nannopterum] harrisi) and the Galipagos Pen- Ballance 1993). Because the majority of seabird guin (Spheniscus mendiculus). Thus, the waters species that live in the tropics are surface-feed- of the Galipagos Islands provide the opportunity ers (Ainley 1977), they often rely on predators to study inshore seabird feeding flocks in one such as dolphin or tuna to drive prey to the sur- area of the tropics that contains sub-surface avi- face (Ashmole and Ashmole 1967, Pitman and an foragers. Ballance 1992). Feeding flocks dependent upon I investigated multispecies seabird feeding dolphin and tuna mainly occur tens to thousands flocks near two islands within the Galipagos Ar- of kilometers from land. For flocks occurring chipelago: the western island, Femandina, and a closer to shore, however, the mechanisms that centrally located island, BartolomC. Two general keep prey close to the surface may differ from questions were investigated: (1) what is the spe- those in the open ocean. The flocks in the Ga- cies composition, frequency, distance offshore, lhpagos are frequent and mainly occur in the in- and duration of the feeding flocks and (2) what shore waters, sometimes within a few feet of effect do the sub-surface seabirds of the Gal& shore. Within these mixed-species flocks, two pagos Islands have on the overall duration of the speciesof sub-surface seabirdsparticipate, mak- feeding flocks? ing it feasible that they play a role in helping METHODS maintain these inshore feeding flocks. Speciesof marine birds that forage by pursuit- STUDY AREA AND DATA COLLECTION diving are restricted to water that has a high This study took place in the Galgpagos Archi- abundance of prey, and thus are generally more pelago, Ecuador during 1994. I spent 58 days at abundant at high latitudes (Ainley 1977, Dia- Cape Douglas, on Femandina Island, and 21 mond 1978). Despite the fact that the Gal&pagos days on Bartolomt (Fig. 1). Data were gathered Islands are within the relatively unproductive between June and August on Femandina and in waters of the tropics, the confluence of currents September on Bartolomk. Sea-surface-tempera- and associated high productivity (Houvenaghel ture (SST) was taken daily at 07:OOfrom an ex- posed section of shore with a mercury thermom- ’ ’ Received26 February1997. Accepted 11 Novem- eter at approximately 50 cm depth. The SST ber 1997. reading that was recorded for the day was the

K’71 278 KYRA L. MILLS

Scale (km)

0 25 50 75 loo

Soundings are in meters

I0 89O FIGURE 1. Bathymetricmap of the GalapagosArchipelago.

average of three separatereadings taken consec- alcohol for subsequentidentification. Two of the utively. samples were taken from feeding flocks near A total of 588 hours on Femandina and 192 Femandina, and one was collected on a beach hours on BartolomC were spent collecting data on BartolomC. In addition, in my daily swims I on feeding flocks. Continuous observations were was able to approachthe feeding seabirdswithin carried out daily from a fixed station on shore a flock and observe specific behaviors of the dif- between 07:OOand 18:O0. For each feeding flock ferent species. observed, the speciescomposition, distance offshore, and duration of the flock were recorded. DATA ANALYSIS On an average day, it was possible to see flocks I define a feeding flock as one that contains 10 that were up to 1 km offshore, although few or more actively feeding seabirdsthat are within flocks occurred at that distance. The distance 3 m from each other. Often, Storm-Petrels offshore of the feeding flock was estimated to (Oceanodroma spp.) participated in feeding the nearest 100 m based upon several landmarks flocks at both study sites, althoughit was difficult of known distance. The duration of the flock was to identify species.Furthermore, because Storm- estimated to the nearest minute by recording Petrels are small and did not congregatein large with a stopwatchthe approximate time the flock numbers at feeding flocks, they were often diffi- formed and the approximate time that the birds cult to spot and count accurately. For these rea- terminated feeding and dispersed. sons, I excluded them from the analyses. Three fish samples from feeding flocks were Data were checkedfor normality and log-trans- collected with a fish net and were preserved in formed when necessary.Results are reported as SEABIRD FLOCKS IN THE GALAPAGOS 279

TABLE 1. Speciesthat participatedin feedingflocks on Fernandinaand Bartolomtfrom Juneto September 1994. Participationfrequency refers to the numberof timesin which each speciesformed part of a flock. Total number of flocks = 277 for Fernandina and 39 for Bartolomt. Values are means F SD.

Total number individuals Mean no. individuaVRock Palticipafion frequency

Common name Fernandina BartolomC Fernandina Bat&m& Fernandina Batiolomt Audubon Shearwater 22,782 160 133 2 185 80 ? 99 62 5 Blue-footed Booby 9,244 190 73 * 99 10 2 10 46 49 Masked Booby 0 5 0 2?1 0 8 Brown Noddy 1.721 188 20 i- 18 15 + 15 31 33 Brown Pelican 1,064 82 11 & 12 523 37 44 GalapagosPenguin 364 108 5&4 7i4 26 39 Magnificent Frigatebird 51 28 525 655 4 13 Flightless Cormorant 13 0 1 2 0.5 0 4 0 GalapagosSea Lion 14 10 453 325 1 8 Bottlenose Dolphin 65 0 11 27 0 2 0 mean ? SD or SE. The software packagesSYS- Audubon Shearwaters are fairly widespread TAT (ver. 5.0) and STATISTICA (ver. 5.0) were throughout the Archipelago, they participated in used to perform statisticaltests, and a critical val- only 5% of the feeding flocks in the waters near ue of P 5 0.05 was acceptedas significant. Stu- Bartolome. In contrast, off the coast of Feman- dent’s t-tests and both univariate (ANOVA) and dina, they were recorded in 62% of the flocks. multivariate analysis of variance (MANOVA) were used to test for differencesbetween groups. FREQUENCY OF FLOCKS AND SST Canonical correlation analyses were used to de- A total of 277 feeding flocks was recorded near termine whether feeding flocks were correlated Femandina, and 39 off Bartolome Island. Flocks with environmental factors (SST), and linear re- occurred more frequently near Femandina (5.2 gressionsand Pearson correlation analyses were -C 6.8 flocks day-‘, n = 51 flocks), with 0.9 used to test for significant associationsbetween flocks hrl, than near Bartolome (2.8 + 2.1 variables. In tests where multiple comparisons flocks day-‘, n = 17), with 0.4 flocks hr’. were made, I used a Bonferroni-adjusted signifi- The mean SST of the water near Femandina cance level of 0.007, based upon the comparison was significantly colder than the water off Bar- of the eight seabird species. tolome (18.9 + l.o”C , n = 54, and 21.6 -C 1.3”C, n = 21, respectively; t,5 = 8.14, P < RESULTS 0.001). Near the Femandina site, both the num- SPECIES COMPOSITION ber of flocks and the total number of birds in Eight species of seabirds participated in the flocks were negatively correlated with SST (I = feeding flocks at both sites. However, no flock -0.6, P < 0.001 and r = -0.7, P < 0.001, comprised more than five species in the area respectively). The relationship between SST and near Fernandina or more than three near Barto- the abundances of the individual species that lome; the mean number of participant species participated in foraging flocks was negative for was 2 for each site (Femandina: 2.1 -C 1.3, n = all speciesat the Femandina site, except Flight- 277, Bartolome: 2.1 ? 1.1, n = 39). Six out of less Cormorants and Magnificent Frigatebirds eight species were common participants in (Fregutu magnijicens), most likely because of flocks at both islands; the Masked Booby (S&a the small sample sizes for these two species. ductylutru) only participated in flocks near Bar- Near Bartolome, there were no significant rela- tolome, and the Flightless Cormorant only par- tionships between SST and the number of flocks ticipated in flocks near Femandina. The most (r = -0.29, P = 0.34), or the number of birds common participants in feeding flocks were Au- in flocks (I = -0.31, P = 0.35). dubon Shearwaters (Puffinus lherminieri) and Blue-footed Boobies (Sula nebouxii) near Fer- DISTANCE FROM SHORE nandina, and Brown Pelicans (Pelecanus occi- Flocks occurred most often close to shore and dentalis), Galapagos Penguins, and Blue-footed decreased in frequency with distance from shore Boobies near Bartolome (Table 1). Even though (Fig. 2). Near Bartolome, 71% of the flocks oc- 280 KYRA L. MILLS

60 , I 1m FERNANDINA •@ BARTOLOME 1

O-50 101-150 201-250 301-350 401-450 501-550 601-650

DISTANCE FROM SHORE (m) FIGURE 2. Frequency distribution of the distance offshore of feeding flocks near Femandina and Bartolome Islands. curred within 60 m of shore (range 2-350 m). The different specieswithin the flocks foraged The flocks near BattolomC occurred on average nonrandomly with respect to the distance from closer to shore than those in the waters off Fer- shore. Flocks of birds where Galapagos Pen- nandina (81.1 ? 93.1 m , n = 31 and 119.5 -t guins, Brown Pelicans, and Brown Noddies (An- 139.2 m, IZ = 267, respectively; r = 0.15, F,,,,, 024sstolidus) were present were significantly = 7.0, P < 0.01). closer to shore than when these specieswere ab- Near Fernandina, there was a significant neg- sent (Table 2). At Bartolome, the trends were ative correlation between the number of species similar to those on Femandina, although not sig- in flocks and the distance offshore (r = -0.26, nificant. Blue-footed Boobies and Audubon P < 0.01, n = 185 flocks), but a significant pos- Shearwatersforaged the farthest offshore at Fer- itive correlation between the total number of nandina, whereas Masked Boobies foraged the birds in the flock and the distance offshore (r = farthest offshore at Bartolome. 0.21, P = 0.001, n = 185). Thus, there was a reduction in species diversity with an increase DURATION OF FLOCKS in the distance offshore of the flocks, even The mean duration of the flocks near Femnndina though flocks increased in size with distance was 5 1.3 + 102.1 min, significantly greater (t,, from shore. Near Bartolome the trend was sim- = 2.1, P < 0.05) than the mean duration of the ilar, although not significant for the number of flocks near Bartolome (20.2 + 21.4 mm). In the species or number of birds. waters near Femandina, about half of the feed-

TABLE 2. Mean distanceoffshore for flocks where GalapagosPenguins, Brown Pelicans, and Brown Noddies were present and absent at the Femandina site. Total number of flocks = 259.

Distance offshore Cm)

GalapagosPenguin 8.5 -+ 75.4 138.5 2 156.9 3.3** Brown Pelican 73.5 ? 82.4 152.0 ? 160.5 5.3** Brown Noddy 94.6 t- 106.2 135.9 ? 154.2 2.8*

*P < 0.01. **p < 0.001. SEABIRD FLOCKS IN THE GALAPAGOS 281

6 I0‘ 7

0 -J 1 2 3 4 5 6 0 ABSENT IPRESENT Number of species FIGURE 4. Duration of feeding flocks (2 SE) with FIGURE 3. Mean (2 SE) duration of feeding flocks GalhpagosPenguins and Flightless Cormorantspresent composedof one to six speciesnear FernandinaIsland. and absent near Fernandina and Bartolomt Islands. Solid bars representBartolomC and the open bars rep- resent Fernandina. ing flocks (53%) lasted 20 min or less. However, SPECIES ASSOCIATIONS 16% of the feeding flocks observed had a duration greater than 1 hr, and several lasted most Certain species of seabirds in the Galapagos of- of a day. The maximum duration of all the flocks ten were found together. Near Fernandina, recorded was 11 hr and occurred at the Feman- Brown Noddies were positively associated with dina site on July 11. penguins (r = 0.31, P < 0.001, n = 268 flocks) Near Femandina, the number of species in a as well as with pelicans (r = 0.59, P < 0.001, flock and the total number of birds in a flock n = 268 flocks). Brown Pelicans were positively both increased with the duration of flocks (num- associatedwith GalLpagosPenguins both at Fer- nandina and BartolomC (Femandina: r = 0.28, ber of species: r = 0.49, P < 0.001, iz = 185 P < 0.001, n = 268; BartolomC: r = 0.34, P < flocks; and number of birds: r = 0.36, P < 0.001, n = 28). In addition, penguins were pos- 0.001, IZ = 185 flocks). Near BartolomC, there itively associatedwith Flightless Cormorants at also was a positive correlation between the num- Femandina (r = 0.27, P < 0.001, iz = 54). Au- ber of birds that formed a flock and the duration dubon Shearwaters and Blue-footed Boobies of the flock (r = 0.50, P < 0.05, n = 16), al- rarely occurred together, despite the fact that though there was no relationship between the these two species were the most numerous in number of speciesand the duration of the flock. feeding flocks. These two speciesco-occurred in Feeding flocks that contained multiple species only 17% of the flocks that I recorded. Indeed, were of greater duration than single species there was a significant negative correlation be- flocks (tlY2 = 5.0, P < 0.01, Fig. 3). tween the presence of these two species (r = Flocks that included diving birds (penguins -0.43, P < 0.01, n = 268). and/or Flightless Cormorants at Femandina and Flocks were significantly larger at Femandina only penguins at BartolomC) were of greater du- when Blue-footed Boobies or Audubon Shear- ration than those that did not contain these spe- waters were present (boobies: tze6= 4.43, P < cies (t,87 = 3.4, P < 0.01, Fig. 4), although the 0.001; shearwaters:tzG6 = 4.42, P < 0.001). results for Bartolomk were not significant. There was a significant negative relationship between PREY AND SEABIRD BEHAVIOR the abundance of Blue-footed Boobies and the Collected fish samples from Femandina were duration of flocks near Femandina (r = -0.15, identified by l? Smith as Pacific Sardines (Sur- P < 0.05, rt = 90), suggesting that large num- dinops sagax), and the fish collected at Barto- bers of Blue-footed Boobies tend to disrupt 1omCwas identified by A. Kitaysky as possibly flocks. a sandlance (Ammodytes spp.). 282 KYRA L. MILLS

Within individual flocks, penguins tended to surface around the two westernmost islands of dive under the school and feed on the way up Femandina and western Isabela. This phenom- to the surface, or picked off fish that were at the enon results in the influx of nutrients into the bottom sides of the school (Mills, unpubl. data). upper water layers, lowering the temperature of Flightless Cormorants approachedthe fish from the surface waters and increasing its productivity the sides, preying on fish in the periphery of the (Houvenaghel 1978, 1984, Feldman 1986). school. The behavior of these two species ap- The spatial variability in oceanographiccon- peared to keep the fish school close to the sur- ditions within the archipelagomay result in a pos- face and clumped, where surface or near-surface sible difference in the type or quantity of prey feeding birds such as Audubon’s Shearwaters, that is available at the two islands in the present Brown Pelicans, and Brown Noddies could ac- study, which may explain the differences ob- cess the fish more easily. Blue-footed Boobies, served in the flocks. There is little information on on the other hand, generally dive in the center pelagic fish communities within the archipelago of the flock (Mills 1996), and presumably in the and the information gathered from the present center of the fish school. In large feeding flocks, study is not sufficient to answer this question. Blue-footed Boobies coordinate their diving, so However, I never observed large schoolsof sar- that they dive simultaneously. This coordinated dines on Bartolome as I did on Femandina. An- activity may stun and confuse the fish momen- other possible explanation of the differences re- tarily, allowing the boobies to feed on them. corded in the feeding flocks may be the different times within the cold season in which the data DISCUSSION were collected, and basedupon this, one may pre- dict flocks with different species compositions. SST AND FEEDING FLOCKS Except for the fact that Flightless Cormorants A greater number, frequency, and duration of were absent from the central island of Bartolome, flocks were observed off Femandina, the island and Masked Boobies were absent from the Fer- with the lower average SST; the number of nandina flocks, the remainder of the specieswere flocks and abundances of the individual species found at both sites and breed in the vicinity of was greater at lower SST at both sites. Seabird both areas(Harris 1982). Even though the species flocks occur mostly when SST is below 23°C that were recorded in the flocks at both islands because low SST is associatedwith higher pro- were similar, the participation frequency within ductivity and a more abundant food supply the flocks differed between islands. The most (Boersma 1974, 1978). In the Galapagos, there striking difference was the participation of Au- is spatial as well as temporal variation in SST dubon Shearwaters. This speciesparticipated in and productivity, either of which, or both, may almost every flock at Femandina, but only par- have influenced the species composition, fre- ticipated in two at Bartolome. It is unclear why quency, distance offshore, and duration of the this was the case, although my observations of flocks recorded in this study. foraging Audubon’s Shearwaters coincide with At both sites there was a general decreasein the more productive areas of the archipelago. SST as the cold seasonprogressed. However, the Based upon the lower SST, the increasein seabird overall SST was colder on Fernandina, and it activity, and the greaternumber and frequency of was coupled with an increase in seabird activity, feeding flocks near Femandina, it seems fair to including an increase in the number and fre- conclude that prey were more abundant at this quency of flocks, migration of penguins into the site than in the area around Bartolome, at least at area, and the onset or increase in the reproduc- the times that I was collecting data. tion of penguins and Flightless Cormorants (Mills 1994). In contrast, no increase in activity THE EFFECT OF DIVERS ON FLOCK was observed on Bartolome during my stay, al- DURATION though the time that I was on Bartolome was When penguins or cormorants were present, the less than my stay at Femandina. flocks lasted longer near Femandina. In the Bar- The spatial variation in sea-surface tempera- tolome area, flocks lasted longer when penguins tures and productivity in the Galapagos Islands were present, although the results were not sig- is the result of the Cromwell Current, which nificant. The GalapagosPenguin population near travels from west to east, and is forced to the Bartolome is relatively low compared to the Fer- SEABIRD FLOCKS IN THE GALAPAGOS 283 nandina population (Mills 1993, Soria et al. other speciesof birds that are unable to dive for 1994, Vargas 1995). On five occasions,feeding their prey. Sub-surface foragers accomplish this flocks that were driven by fish such as tuna (un- by diving at the periphery of the flock and ap- identified species) were observed near Bartolo- proaching prey from the bottom or edges of the me. It is possible that because of the smaller school, as my observations of penguins and cor- number of sub-surface avian foragers found in morants suggest (Mills, unpubl. data). In near- this area, flocks more often are driven and main- shore waters, flocks may form as a result of the tained by sub-surface predators such as tuna, concentration of prey due to physical effects which would be similar to the driving force of such as fronts, internal waves, and tidal currents flocks in the open ocean and away from shore. in combination with shallow water, or by bio- I also observed feeding flocks around southern logical factors such as sub-surface animals driv- Isabela and Femandina islands that were initi- ing prey to the surface. In certain areas of the ated by Sierra mackerel (Scomberomussierra). Galapagos, these neat-shoreflocks, once formed, These mackerel-based flocks were seen while I are maintained by pursuit-diving species.Away was traveling between islands, and the flocks from shore, where diving seabirds are rare or were farther offshore than those flocks recorded absent, surface or near-surface foraging seabirds in the present study. rely on other methods of keeping prey close to Diving animals play important roles in main- the surface. taining multispecies feeding flocks in many oceans of the world. In the eastern tropical Pa- SPECIES INTERACTIONS cific, a high percentageof seabird feeding flocks Flocks form where there is an abundance of food are initiated and maintained by sub-surfacepred- available to predators (Sealy 1973, Haney and ators, specifically dolphin (Stenella attenuata McGillivary 1985). Once one or several birds and S. longirostris) and yellowfin tuna (Thunnus begin actively feeding, other birds that are in the albacares), which drive smaller fish to the sur- vicinity approach that area and begin feeding as face (Au and Pitman 1986, Ballance 1993). Pit- well, which in turn attracts other birds from far- man and Ballance (1992) found that 76% of the ther away (Sealy 1973, Hoffman et al. 1981, Ha- Parkinson’s Petrels (Procellaria parkinsoni) that ney et al. 1992). Seabirds continue to feed until they observed were associatedwith feeding dol- they deplete the food resource, become satiated, phins. In addition, Sooty Terns (Sternafuscata), or until the prey are able to escape. Visual re- because they are unable to feed deeper than a cruitment is an important factor in flock forma- few centimeters below the surface of the water, tion. This importance is clearly revealed in the also depend upon prey made available to them positive relationship found in the present study by sub-surface predators (Au and Pitman 1986). between the number of species, the number of In the Bering Sea, in 67% of the incidences individuals, and the duration of the flock. Thus, where feeding gray whales (Eschrichtius robus- flocks that last longer than others have a better tus) created mud plumes, seabirdsforaged in the chance of attracting a greater number of species plumes (Obst and Hunt 1990). Several diving and a greater number of birds. Hoffman et al. seabirds, such as species of Antarctic and Sub- (1981) used the name of “catalysts” to describe Antarctic penguins (Harrison et al. 1991), Mag- those specieswhose active foraging is very con- ellanic Penguins Spheniscusmagellanicus (Jehl spicuousand is used by other birds as indicators 1974, Boswall and MacIver 1975), Murres Uria of areaswith a high abundance of prey, therefore spp. (Chilton and Sealy 1987, Hunt et al. 1988), allowing a rapid development of a flock. I found and Marbled Murrelets Brachyramphus mar- that Audubon Shearwatersand Blue-footed Boo- moratus (Mahon et al. 1992) have been noted to bies were the catalystsfor the Femandina hocks, actually make prey available to other species and Brown Pelicans and penguins were the cat- (usually surface-feeding seabirds) by driving alysts of flocks in the vicinity of Bartolome. prey to the surface, keeping prey close to the Shearwaters and Boobies are very conspicuous surface, or by inadvertently injuring or killing when they are feeding on a prey patch. However, prey that can be fed upon by birds at the surface. foraging penguins are not highly visible and it If the presenceof diving speciesincreases the may seem odd that they would act as catalysts. overall duration of flocks, then this suggeststhat This may be explained by the fact that penguins divers actually help in making prey available to are highly associatedwith pelicans, whose feed- 284 KYRA L. MILLS ing is highly visible. In my study, it was almost aging with boobies, they were feeding on zoo- always the case that flocks that contained pen- plankton, although this has not been determined. guins also contained pelicans. “Suppressors” are those species that disrupt ACKNOWLEDGMENTS the snatial structure of a flock because of their I thank several reviewers for helpful commentson ear- feedihg habits (Hoffman et al. 1981). In the Ga- lier drafts of this manuscript:Priyanga Amarasekare, Lisa Ballance, Alistair Cullum, Brad Hawkins, George lipagos, Blue-footed Boobies, which are plunge- L. Hunt Jr., Alexander Kitaysky, Jen Zamon, and one divers, may act as suppressors, given their ten- anonymousreviewer. I am-grateful to Brad Hawkins dency to dive in the center of a fish school. This and Alexander Kitaysky for their help with statistical type of feeding method may cause prey to dis- analyses.Paul Smith (Scripps Institution of Oceanog- raphy) and Alexander Kitaysky (University of Califor- perse and therefore shorten the longevity of the nia-Irvine) helped with fish identification. I gratefully flock. Likewise, Bayer (1983) hypothesized that acknowledge the help of my field assistants,M6nica the feeding flocks that he observed in Alaska Patricia Caivopifia, Manuel Feijbo, Alex Gomez Coel- were of short duration because the feeding ac- lo, Loren Knopper, Paula Kohn, and Norma Yanez Caluachi. Logistical support was provided by the tivity of the plunging Black-legged Kittiwakes Charles Darwin ResearchStation (CDRS), the Charles (Larus triductylu) dispersed the prey. Hoffman D arwin Foundation. and the Gal&JagosNational Park et al. (1981)\ , also found that both the Short-tailed Service (GNP%. TAME provided a discountedairline (PufJinus tenuirostris) and Sooty Shearwaters ticket to’the Islands. Many thanks to Arturo Izurieta and Felipe Cruz (ex-director and ex-subdirector, re- (P. griseus), which pursuit-plunge, dispersed spectivelv of the GNPS), who made the work on the prey. islands n&h easier by ;he loan of a dinghy and a trip Species are attracted to a common resource to BartolomCon the GNPS yacht. This work was fund- and therefore, as Veit (1995) notes, positive as- ed by the Department of Ecology and Evolutionary Biology at the University of California, Irvine, grants sociations between different species are expect- from the Los Angeles Audubon Society, the Chicago ed. However, the nature of the interactions Zoological Society, the Penguin Fund, Georgina and among species may be useful in predicting August0 Cmz from the RN Beagle III, and personal which species are most often found together in donationsby George L. Hunt Jr., and Nick and Dolores Mills. Permission to do research on the islands was flocks and which are not (in the case of a neg- provided by the GalapagosNational Park Service. ative relationship). Where feeding niches over- lap, certain species may actually avoid each oth- LITERATURE CITED er (but see Veit 1995). Ashmole (1968) found AINLEY,D. G. 1977. Feeding methods in seabirds:a that around Christmas Island, Anous stolidus fed comparison of polar and tropical nesting com- closer to land than Sterna fuscata, which flew munities in the EasternPacific Ocean,p. 669-685. over flocks of A. stolidus to forage further off- In G. A. Llano [ed.], Adaptationswithin Antarctic ecosystems. Smithson. Inst. Press, Washington shore, even though they feed on similar-sized DC. prey. The only negative relationship encountered ASHMOLE,N. P 1968. Body size, prey size and eco- in the present study was between Blue-footed logical segregationin five sympatrictropical terns Boobies and Audubon Shearwaters. This sug- (Aves: Laridae). Syst. Zool. 3:292-304. ASHMOLE,N. P, AND M. J. ASHMOLE.1967. Compar- gests that these two species actively avoid each ative feeding ecology of sea birds of a tropical other, although the reason for this is unclear. Per- oceanic island. Peab;;dy Mus. Nat. Hist. Buli 24. haps Audubon Shearwaters tend to avoid areas Au. D. W. K.. AND R. L. PITMAN. 1986. Seabird in- where Blue-footed Boobies are feeding because teractions with dolphins and tuna in the eastern tropical Pacific. Condor 88:304-317. of their prey-dispersing tendencies. Another pos- BALLANCE,L. T. 1993. Community ecology and flight sible explanation is that they may be feeding on energetics in tropical seabirdsof the Eastern Pa- different prey. Boersma (1974) and Harris cific: energetic correlatesof guild structure.Ph.D. (1977) both report that Audubon Shearwaters in diss., Univ. California, Los Angeles, CA. BAYER,R. D. 1983. Black-legged Kittiwake feeding the Galipagos feed on zooplankton and Blue- flocks in Alaska: selfish/reciprocal altruisti; footed Boobies feed on medium-sized fish (Har- flocks? J. Field Omithol. 54:19&199. ris 1977). However, on several occasions on Fer- BOERSMA,I? D. 1974. The Galapagos Penguin: ad- nandina I observed shearwaters feeding on sar- aptationsfor life in an unPred&ble envgonment. Ph.D. diss., The Ohio State Univ., Columbus,OH. dines as did other species in the feeding flocks, BOERSMA,F? D. 1978. Breeding patternsof Galhpagos including Blue-footed Boobies. Perhaps on those Penguins as an indicator of oceanographiccondi- occasions when shearwaters were not found for- tions. Science 200:1481-1483. SEABIRD FLOCKS IN THE GALAPAGOS 285

BOSWALL,J., AND D. MACIVER. 1975. The Magellanic St. Matthew Island, Bering Sea. Auk 105:345- Penguin Spheniscusmagellanicus, p. 271-306. In 349. B. Stonehouse [ed.], The biology of penguins. JEHL, J. R., JR. 1974. The distribution and ecology of MacMillen Press, London. marine birds over the continental shelf of Argen- CHILTON,G., AND S. G. SEALY. 1987. Speciesroles in tina in winter. San Diego Sot. Nat. Hist. Trans. mixed-species feeding flocks of seabirds.J. Field 17:217-234. Omithol. 58:45&463. MAHON,T E., G. W. KAISER,AND A. E. BURGER.1992. DIAMOND,A. W. 1978. Feeding strategiesand popu- The role of Marbled Murrelets in mixed-species lation size in tropical seabirds.Am. Nat. 112:215- feeding flocks in British Columbia. Wilson Bull. 223. 104:738-743. DUFFY,D. C. 1983. The foraging ecology of Peruvian MILLS, K. L. 1993. Report of the 1993 annual census seabirds.Auk 100:800-810. of the Galapagos Penguin (Spheniscusmendicu- FELDMAN,G. C. 1986. Patternsof phytoplanktonpro- lus) and Flightless Cormorant (Nannopterumhar- duction around the Galapagos~sl&&, p. 77-iO6. risi) in the &l&pagos Archipelago. Report to the In M. J. Bowman. C. M. Yentsch. and W. T Pe- Charles Darwin Res. Stat. and the Gal&JaeosNatl. terson reds.], Lecture notes on coastal and estua- Park Serv., Santa Cruz, Galapagos_ - I&n&. rine studies:Vol. 17. Springer-Verlag,New York. MILLS, K. L. 1994. Field seasonreport. Rep. submit- HANEY.J. C.. K. M. FRISTRUP.AND D. S. LEE. 1992. ted to the Charles Darwin Res. Sta. and the Ga- Geometry of visual recruitment by seabirds to lapagos Natl. Park Serv., Santa Cruz, Galhpagos ephemeral foraging flocks. Omis Stand. 23:49- Islands. 62. MILLS, K. L. 1996. Seabird foraging behavior in the HANEY,J. C., AND F! A. MCGILLIVARY.1985. Flocks inshore waters of the Galapagos Islands. M.Sc. of Cory’s Shearwaters(Calonectris diomedea)at thesis. Univ. California, Irvine, CA. Gulf Stream fronts. Wilson Bull. 97: 191-200. OBST, B. S., AND G. L. HUNT JR. 1990. Marine birds HARRIS,M. P. 1977. Comparative ecology of seabirds feed at gray whale mud plumes in the Bering Sea. in the Galapagos Archipelago, p. 65-76. In B. Auk 107:678-688. Stonehouse and C. Perrins [eds.], Evolutionary PITMAN,R. L., AND L. T BALLANCE.1992. Parkinson’s ecology. Univ. Park Press, Baltimore, MD. Petrel distribution and foraging ecology in the HARRIS,M. 1982. A field guide to the birds of Gal& easternPacific: aspectsof an exclusive feeding re- pagos. 2nd ed. Radavian Press, Reading, UK. lationship with dolphins. Condor 94:825-835. HARRISON,N. M., M. J. WHITEHOUSE,D. HEINNEMAN, SEALY,S. G. 1973. Interspecific feeding flocks of ma- F?A. PRINCE,G. L. HUNT JR., AND R. R. VEIT. rine birds off British Columbia. Auk 90:796-802. 1991. Observationsof multispeciesseabird flocks SORIA,M., K. L. MILLS, AND L. CAYOT. 1994. Report around South Georgia. Auk 108:801-810. of the 1994 annual censusof the GalapagosPen- HOFFMAN,W., D. HEINEMANN,AND J. A. WIENS. 1981. guin (Spheniscusmendiculus) and Flightless Cor- The ecology of seabird feeding flocks in Alaska. morant (Nunnopterumharrisi) in the Gahipagos Auk 98~437-456. Archipelago. Report to the Charles Darwin Res. HOUVENAGHEL,G. T 1978. Oceanographicconditions Stat. and the Galapagos Natl. Park Serv., Santa in the Galapagos archipelago and their relation- Cruz, GalapagosIslands. ship with life on the islands, p. 18l-200. In R. VARGAS,H. 1995. Report of the 1995 annual census Boje and M. Tomczak [eds.], Upwelling ecosys- of the Galkpagos Penguin (Spheniscusmendicu- tems. Springer-Verlag.New York. lus) and Flightless Cormorant (Cumpsohaelius HOUVENAGHEL,G. T. 1984. Oceanographicsetting of [Nannopterum]harrisi) in the Galapagos Archi- the GalapagosIslands, p. 43-54. In R. Perry [ed.], pelago. Report to the Charles Darwin Res. Stat. Galhpagos.Pergamon Press, Oxford. and the Galapagos Natl. Park Serv., Santa Cruz, HUNT, G. L., JR., N. M. HARRISON,W. M. HAMNER, GalgpagosIslands. AND B. S. OBST. 1988. Observationsof a mixed- VEIT, R. R. 1995. Pelagic communitiesof seabirdsin speciesflock of birds foraging on euphausiidsnear the South Atlantic Ocean. Ibis 137:1-10.