Revista Chilena de Historia Natural 57:47-57, 1984 Wilson's Phalarope in the Central Andes and its Interaction with the Chilean *

Pollito de mar tricolor en los Andes Centrales y su interacci6n con el flamenco chilena

STUART H. HURLBERT¹,MATILDE LOPEZ 2 and JAMES 0. KEITH 3

1 Departament of Biology, San Diego State University, San Diego, California 92182, USA. 2 Departamento de Ciencias Naturales, Universidad de Antofagasta, Antofagasta, Chile. 3 1771 South Welch Circle, Lakewood, Colorado 80228, USA.

ABSTRACT

More than 500,000 Wilson's phalaropes (Phalaropus tricolor) have been observed during the austral summer on shallow, saline lakes of the Bolivian puna. This region, with adjacent parts of Chile and , thus appears to be the major wintering grounds of the species. The lakes on which it is found in greatest numbers are simultaneously inhabited by large numbers of the Chilean flamingo (Phoeni- copterus chilensis) and populations of a few invertebrates on which both birds feed: brine shrimp (Artemia salina), copepods (Boecke/la poopoensis), chironomid midges (Paratrichocladius sp.), and brine flies (Ephydridae). As P. cbilensis feed, P. tricolor individuals sometimes swim rapidly around them and feed on invertebrates that these larger companions stir up from the lake bottom. P. tricolor ignores the other two flamingo species (Phoenicoparrus andinus, P. jamesi) present on the lakes; these feed on and move about less rapidly than doesP. cbilensis. Keywords: altiplano, saline lakes, zooplankton, , Chile.

RESUMEN

Más de 500.000 Pollitos de mar tricolor (Phalaropus tricolor) han sido observados durante el verano austral en las someras lagunas salinas de la puna boliviana. Al parecer, esta region, en conjunto con re- giones adyacentes de Chile y Argentina, constituye el área principal de invernaci6n para la especie. Las lagunas donde esta ave se encuentra con mayor abundancia están pobladas tam bién por grandes numeros del Flamenco chileno (Phoenicopterus chilensis) y por densas poblaciones de unos pocos invertebrados utilizados por las dos aves como alimento: branqui6podos (Artemia salina), copepodos (Boeckella poopoensis), quiron6midos (Paratrichocladius sp.), y moscas de playa (Ephydridae). Cuando los indivi- duos de P. chilensis están alimentándose activamente, a veces unos ejemplares de P. tricolor nadan rapidamente alrededor de ellos y aprovechan los invertebrados que los primeros revuelven desde el fondo de la laguna. P. tricolor ignora las otras dos especies de flamenco (Phoenicoparrus andinus, P. jamesi) que se hallan en las lagunas; estas especies se alimentan de diatomeas y se mueven de un sitio a otro más lentamente que P. chilensis. Palabras claves: altiplano, lagos salinos, zooplancton, Bolivia, Chile.

"The mystery which surrounds the winter quarters of pacidae. We here report observations indi- the Phalarope has always been an interesting study ... " cating that it is one of the most abundant MEINERTZHAGEN (1925) birds in the Central Andes, though there are very few published records of its pre- INTRODUCTION sence there. We also describe its close Wilson's phalarope or the Pollito de Mar association with the Chilean flamingo Tricolor (Phalaropus tricolor) (Viellot) (Phoenicopterus chilensis Molina) in this is an unusual member of the family Scolo- region.

* Invited paper Offprints requests to: S. Hurlbert or M. López.

(Received 16 August 1983. Accepted 10 October 1983). 48 HURLBERT ET AL.

P. tricolor typically nests in herbaceous Jehl, pers. comm.) it is apparent that the vegetation around lakes and marshes in species main wintering areas have gone southern Canada and northwestern United undetected or at least unreported. States (Bent 1927, Palmer 1967). It has long been suspected to overwinter METHODS in South America and, indeed, there are Our observations of Wilson's phalaropes numerous records of its occurrence there have been made during a number of expe- during the austral summer (records sum- ditions to the Central Andes since 1971. marized by J. Jehl, in manuscript). However, The primary purpose of these expeditions there are no published records of the has been a comparative study of altiplano species having been seen there in large lakes and the three flamingo species (Phoe- numbers. The largest flock seen by Wet- nicopterus chilensis, Phoenicoparrus andi- more ( 1926) consisted of 40 individuals nus (Philippi), Phoenicoparrus jamesi (Lago Epiquen, Argentina). For Buenos (Sclater) inhabiting them (Hurlbert 197 8, Aires Province, Olrog ( 1967) reported 1981, 1982, Hurlbert and Chang 1983, flocks of up to 28 individuals and Myers Hurlbert and Keith 1979; Hurlbert, Lo- and Myers ( 1979) likewise noted that pez and Stein, unpub. data) and analysis "Flock sizes were generally small, but of the structure and origin of massive ranged to > 50 birds". But in the Argenti- blocks of ancient freshwater ice present ne interior, 30,000 - 50,000 or more in- in a few of the highest salt lakes in the dividuals have been observed on both region (Hurlbert and Chang, unpub. data). Laguna Pozuelos (22°00' S, 66°00' W; On every expedition since 1974 we have ca. 3500 m) and Mar Chiquita (30° 40' S, made a detailed census of the 62° 30' W) (C. Olrog, pers. comm.), and present on each lake visited. For other "hundreds" were observed in February bird species, such as P. tricolor, we have 1972 on "a small lake about 10 km north kept less systematic records and usually of Abra Pampa", just south of Laguna Po- have estimated numbers only when these zuelos (M.P. Kahl, pers. comm.). The were great or when detailed behavioral largest flock recorded for Chile was 20 observations were made. Censuses have birds (Johnson 1965), but M. Sallaberry been made using 1OX binoculars and 15- (pers. comm.) has recently observed more 60X zoom spotting scopes. than 200 at Lago Chungara ( 18° 15' S, Quantitative zooplankton samples have 69°09' W; 4520 m). For coastal lagoons been collected from all lakes visited. The and pools of southern Peru, P. tricolor sampling procedure consisted of walking has been reported as "common on passage out into the lake with a 3.5 liter saucepan with some remaining in the district through- ( cacerola) and plunging in into the always out the austral midsummer months" shallow water 5-20 times, each time pour- (Hughes 1970). In the Peruvian Andes, ing the collected water through a zoo- "hundreds of thousands" of phalaropes plankton net with 50-60 μm mesh apertu- were observed on Lago J unin ( 11 o 00' S, res. Samples were preserved in 5-l 0% for- 76°20', W; 4080 m) in early October maldehyde. 19 77 (J. Fjeldsa, pers. comm.) and on Laguna Las Salinas (16°22' S, 71°09' W; RESULTS 4295 m) in mid-April 1973 (M.P. Kahl, pers. comm.); and there have been several Phalarope Abundance and Distribution sightings of small numbers of phalaropes on lakes to the north and west of Lago Ti- Our first observations of P. tricolor in the ticaca, as well as on Lago Titicaca itself altiplano were made in Salar de Carcote (J. Fjeldsa, pers. comm., Holmes 1939). (21°20' S, 68°21' W), Chile, during visits in A strong tendency to co-occur has been December 1971 and March and December observed for P. tricolor and P. chilensis 1972. In the numerous, small ( < 0.5 2 on these lakes (Fjeldsa 1981 ). All other km ) saline ponds around the margin records for South America are for small of this salar, we observed P. tricolor in ( < 30) or unstated numbers of individuals. numbers ranging from 1 to 14 per pond. Since the number of Wilson's Phalaropes Larger flocks of P. tricolor were not migrating southward from North America observed until we began study of lakes at the end of each breeding season is on of the Bolivian puna (Fig. 1) in 1975. the order of 1,000,000 - 3,000,000 (J. In Table 1 we list all of our records for WILSON'S PHALAROPE IN THE CENTRAL ANDES 49

Canapa of Hedionda THE \ Ramaditas CENTRAL

\ Grandes kilometers 25 ! J Laguna Khara

) ' Laguna

J i

\ Laguna Loromayu -Laguna coruto \ Rio ) de Chalviri Negras \ Laguna

(' Laguna Kalina-' / i

.. ____ .

de Aguas j de

68o /

Fig. 1: Location of lakes in the Central Andes where large num- bers of Wilson's phalarope occur during the austral summer. Ubicaci6n de los lagos en los Andes centrales donde se encuentran grandes numeros de P. tricolor durante el verano austral.

P. tricolor flocks exceeding I 00 birds. but did not keep written records of the Smaller flocks have been observed during observations. the austral summer in Bolivia on Laguna Our estimates of the sizes of the larger Totoral (22° 32' S, 67° 17' W) and in Salar P. tricolor flocks (Table I) are very approx- de Chalviri (Laguna Polques, Laguna Nor- imate, are given verbatim from our te, Laguna Puripica Chica, Laguna Herre- field notes, and may in some cases be in ra - see Hurlbert and Chang 1983, for error by 50 percent or more. For small exact locations; 22º33' S, 67° 33' W), birds in large flocks on large and briefly in Chile on Laguna Lejia (23° 30' S, 60° visited lakes, more accurate estimates are 42' W) and in Laguna Salada (23° 41' S, not possible. Our census data for the much 68°08' W, Salar de Atacama), and in larger flamingos are more reliable and Peru (November 1976) on Laguna Chaccas rarely would be in error by more than (15°25' S, 70° 12' W) and Lago Titicaca I 0-20 percent. (Puna Bay, 15°45' S, 69°50' W). We also The number of P. tricolor that over- have seen P. tricolor in small numbers winter in the southwestern tip of Bolivia on several other lakes in the Bolivian puna is best indicated by our observations in TABLE 1 Records of Wilson's phalarope on Bolivian lakes, with data on flamingo numbers and other features of the lakes Registros de P. tricolor en lagos bolivianos, con datos sobre numeros de flamencos y otras caracteristicas de los lagos.

Lake Elevation Water No. of birds on lake (latitudeS above sea surface a P. P. P. Unidentified P. longitude W) level (m) (km 2 ) (O/oo) Date chilensis andinus jamesi flamingos tricolor

Laguna Cañapa 14 Feb. 1979 55 2 4. least (21 º30'' 68º01 ') 23 Feb.1979 29 1 6 Laguna Chulluncani 19-69 12 Dec. 1976 354 47 (no data) (21º32', 67º52') 22 Feb. 1979 1161 56 86 Laguna Ramaditas 4117 28-38 17 Dec. 1975 15 (21 º38'' 68º05') 12 Dec. 1976 67 29 0 Laguna Khar Kkota 4112 17 Dec. 1975 102 15 (21035', 68º04') Laguna Hedionda II 4121 12 Feb.1979 4154 45 (21 º34', 68º03') Laguna Colorada 4278 16 Dec.1975 46 2 5965 8987 67º47') 24 Jan. 1979 634b 228b 7648b Laguna Loromayu 7.5 94 18 Feb.1979 hundreds (22º18'' 67º13') of Rio Puntas Negras 14e 18 Feb. 1979 14 1 (22º23', 67º04') Laguna Kalina 16 43-65 1 Feb. 1979 1675 958 6934 few hundred

Laguna VerdeC 4315 49-55 13 Dec. 1975 588 12 (22º48', 11 Dec.1976 734 35 6 Jan. 1983 92 119 4 a. Determined with an American Optical hand refractometer (Model10419). Ranges are for samples taken in different locations and/or different years. b. These data obtained on January 21. c. Data (including area) are for eastern part of lake only. d. Only the northwestern most corner of the lake was censused for P. tricolor. e. Value given is actually for Laguna Chojllas from which this river flows. WILSON'S PHALAROPE IN THE CENTRAL ANDES 51

the first half of February 1979 (Table of the water column, leading, at least I). At that time three lakes -Hedionda temporarily, a planktonic existence. II, Loromayu and Kalina- were each esti- The zooplankton data illustrate very mated to have I 00,000 or more P. trico- clearly that only a few taxa are present lor on them. The total number present in any one lake,. a condition typical of on the lakes examined at this time was salt lakes. As each sample represents a on the order of 500,000 to 1 ,000,000, single location in the lake, the calculated reasonably assuming that we were not absolute abundances (Table 2) should simply censusing a single mobile flock be regarded as only roughly indicative moving from lake to lake. Whether large of true lake-wide densities. numbers of P. tricolor were present at this At all lakes where P. tricolor has been time at any of the unvisited lakes in this observed in large numbers the zooplankton region or in other parts of the altiplano is strongly dominated either by calanoid is not known. copepods (Boeckella poopoensis Marsh) Laguna Kalina was where we first saw or by brine shrimp (Artemia salina Leach) P. tricolor in tremendous numbers. In our (Table 2). The greatest numbers of P. tri- field notes we recorded that "phalaropes color were recorded for lakes with Arte- are everywhere in flocks of thousands mia -Laguna Hedionda II, Laguna Loro- and tens of thousand" bobbing in the open mayu and Laguna Kalina. No lake in lake and resting on mudflats and along the set has both Boeckella and Artemia. the shoreline. At one point along the Salinity is a fairly good predictor of which western shoreline we observed one mudflat will be present, Artemia occurring at the that was dark gray, unlike the whitish higher . calcium carbonate mudflats seen elsewhere After Artemia and Boeckella, the most in the lake. A moment later this darker important prey for the P. tricolor in this "mudflat" self-levitated and dispersed over region may be midges (Chironomidae) the lake: it was a single flock of roughly and brine flies (Ephydridae). During the 30,000 P. tricolor. austral summer chironomids of the genus Paratrichocladius are found in great num- The physical characteristics of the lakes bers at two lakes -Laguna Verde and hosting large numbers of P. tricolor are Laguna Loromayu. Our plankton data extreme. The lakes occur at very high probably underrepresent the real abundan- elevations, are very saline and very shallow ce of their larvae. The adult chironomids (Table 1). Except possibly for Laguna Ver- swarm abundantly around the margins de and Laguna Loromayu, the deeper of these same lakes and, when the wind parts of which we have not inspected, comes up in the afternoon, seek shelter all the lakes have mean depths of less under rocks on the shoreline and vegeta- than 0.3 m. Their ionic compositions tion-free pampa surrounding the lakes. are, in all cases, of the sodium sulfato- However, both adults and pupae may be chloride type (unpubl. data). most available to phalaropes at the water During the summer the lakes contain surface on days of mass emergence (J. dense populations of invertebrates. This Fjeldsa, pers. comm.). Phalarope excrement food source is undoubtedly the principal on the shoreline of Laguna Verde on De- feature attracting phalaropes to these cember 13, 1975 contained remains of lakes. Summer zooplankton abundance both adult and immature chironomids has been determined at one time or another in abundance, as well as lesser quantities for all lakes, though not always at the same of copepod remains. time or year that we recorded phalarope Larvae and pupae of ephydrids have numbers (Table 2). In such shallow lakes, never been found in our plankton samples the distinction between zooplankton and and only occasionally in qualitative samples benthos is somewhat artificial; during of the benthos. However, large swarms the afternoon, when there is almost always of adults have been observed at most a strong wind, crustacean zooplankters of the salt lakes of the region, so the rarity are often observed resting on or holding of the aquatic stages is probably only onto the lake bottom. On the other hand, apparent, a reflection of patchy within- typically benthic organisms such as chiro- lake distributions and a preference for nomid larvae and microscopic nematodes firm substrates, which we did not sample. frequently are found in the upper part Larvae, pupae and adults of ephydrids 52 HURLBERT ET AL.

TABLE 2 Zooplankton abundance in lakes hosting large numbers of Wilson's phalarope in the austral summer Abundancia de zooplanctontes en Iagunas con grandes mimeros de P. tricolor durante el verano austral

Number per liter Lake and Boeckella poopoensis Artemia Date copepodids nauplii salina Other invertebrates

Laguna Cañapa 23 Feb. 1979 45 Laguna Chulluncani 12 Dec. 1976 269 23 Feb. 1979 16.5 6.2 Branchinecta sp. (2.2)a Laguna Ramaditas 12 Dec. 1976 36 3 23 Feb. 1979 nematodes (1.8), harpacticoid copepods (1.8) Laguna Khar Kkota 17 Dec. 1975 0 0 chironomid larvae Laguna Hedionda II 23 Feb. 1979 nematodes (24), harpacticoids Laguna Colorada 27 Nov. 1977 1.3 Laguna Loromayu 1 Dec. 1977 21.4 chironomid larvae (3.1) 18 Feb. 1979 Rio Puntas Negrasb 18 Feb. 1979 Laguna Kalina 2 Dec. 1977 124 nematodes (6. 2) 3 Feb. 1979 nematodes (21.6) Laguna Verde 11 Dec. 1976 46 cyclopoid copepods (27), nematodes (48), chironomids (1.1) 3 Dec. 1977 62.4 22.3 a. Possibly Branchinec ta palustris Biraben 1946 (R. Hartland- Rowe, pers. comm.). b. Quantitative sample was not taken. have been found in abundance in P. trico- 2), as this group is rarely seen in the plank- lor stomachs elsewhere (Bent 1927; J. ton. They may represent benthic forms Jehl, pers. comm.). swept up into the water column by turbu- Branchiopods in the genus Branchinecta lence. The observed nematodes were have been associated with large numbers very small (mostly 200-2000 μm long by of P. tricolor only once -in Laguna Chu- 10-40 μm in diameter) and it is unlikely lluncani at a time when its salinity was that P. tricolor fed on them. only 19° /oo (22 Feb. 1979; Table 2). How- ever, Branchinecta spp. are often the domi- Phalarope-Flamingo Relations nant form present in the small temporary rainwater pools that form in January or P. tricolor is strongly associated with the February and last for a few weeks to a Chilean flamingo (P. chilensis) during few months in wet years. We observed the austral summer both distributionally many such pools only in January and and behaviorally. In the puna region February 1979 but saw no P. tricolor (roughly the Andes between 21 o and 27°) on them. These rainwater pools develop of Bolivia and Chile, the two species occur more abundantly in the rainier central predominantly on the same lakes, as is and northern portions of the altiplano. also the case in the Peruvian altiplano The abundance of nematodes in many (Fjeldsa 1981). The three lakes -Hedionda of the P. tricolor lakes was striking (Table II, Loromayu and Kalina- where P. trico- WILSON'S PHALAROPE IN THE CENTRAL ANDES 53

lor has been recorded in greatest abundance of the feet of P. chilensis. The P. tricolor are the same three lakes where we have re- did not associate at all with the 104 P. corded P. chilensis in greatest numbers andinus or 54 P. jamesi which were present in the puna, though we have censused along with the 197 P. chilensis. flamingos 2-6 times not only for the lakes The numerous P. tricolor on Laguna listed in Table I but also for about 60 Hedionda on February 12 were feeding other lakes in this region (Hurlbert 1978, in close association with the P. chilensis 1981, and unpubl. data; Hurlbert and there, ignoring the P. andinus and P. ja- Keith 1979). Laguna Chulluncani and La- mesi present (Table 1). At Laguna Cafiapa guna Colorada also commonly have large on February 14 we began observing the numbers of both P. tricolor and P. chi- small P. chilensis flock (Table 1) starting lensis present. The rich food supply of early in the morning while they were still invertebrates that makes the lakes of resting, mostly on one leg with heads Table 1 favorable wintering grouns for crooked or tucked under wings, in a group P. tricolor also makes them favorable in the center of the lake. At 0610 hours nesting sites for P. chilensis, which feeds (air temperature = -3°C; sun hit lake on the same aquatic insects and crustaceans at 0655 hours), we observed that "about (Allen 1956, Rooth 1965, Hurlbert, Lo- 60 phalaropes have flown in and swin pez, and Stein, unpub. data). We have around flamingos as if waiting for them observed P. chilensis nesting at Laguna to wake up and feed"; later interactions Colorada, Laguna Loromayu, and Laguna were not recorded. Kalina, though most of the eggs laid are Our most detailed observations on the harvested by the local inhabitants (Hurl- association were made at Laguna Chullun- bert and Keith, unpubl. data). P. chilensis cani on February 22 (Table 1). All three nested at Laguna Verde in the past and flamingo species were present, but again occasionally still attempts to do so, but P. tricolor associated only with actively always unsuccessfully due to visits of the feeding P. chilensis. In the morning, almost eggers. every P. chilensis individual was closely P. tricolor abundance shows weak or accompanied by 1-3 P. tricolor (Table 3), no correlation with the abundances of rapidly swimming between and around the other two flamingos, P. andinus and the flamingo's legs. P. chilensis individuals P. jamesi, on the puna lakes. Though which were resting or standing still lacked P. andinus and P. tricolor sometimes these little companions. In the afternoon, abundantly co-occur on some lakes (e.g. the association broke down somewhat Hedionda II, Kalina: Table 1), P. andinus (Table 3). The light southwest wind, which has been observed in large numbers on began about 1030 hours, may have stirred many puna lakes where P. tricolor was the water column sufficiently that P. uncommon or absent (e.g. in Bolivia, tricolor did not have to depend so much Laguna Polques, Laguna Puripica Chica, on the movements of P. chilensis for Salar de Pastos Grandes, and, in Chile, raising the Boeckella and/or Branchinecta Laguna Lejia: Hurlbert 1981 ). Both P. to the surface. andinus and P. jamesi feed primarily on di- Only on one occasion was P. tricolor atoms (Hurlbert, Lopez and Stein, unpubl. observed to associate with one of the data). other flamingo species. At Laguna Puripica A more intimate association between Chica on February 17 at 1920 hours there P. tricolor and P. chilensis was observed were 12 P. chilensis, 10 P. jamesi, and on Laguna Lejia in December 1975 and 1005 P. andinus present and we recorded on several of the smaller Bolivian lakes "many phalaropes (number unrecorded) in February 1979. At Laguna Lejia we feeding along with the P. andinus -but observed phalaropes following Chilean fla- not in as frenzied fashion as when with mingos, apparently picking up organisms P. chilensis". This lake is only slightly stirred up by them. At Laguna Herrera saline ( 13%o) and contains corixids and on February 5 we noted P. tricolor indi- amphipods as well as copepods. viduals (numbers unrecorded) clustering The preferential association of P. trico- around actively feeding P. chilensis and lor with P. chilensis is almost certainly pecking at the water surface, presumably due to the fact that this flamingo walks feeding on copepods brought to the surface much more rapidly and, consequently, by the rapid stomping or treading motion stirs up the water and lake bottom more 54 HURLBERT ET AL.

than do P. andinus and P. jamesi. We had at one spot. ln certain situations P. chi- noted and quantified this difference se- lensis will remain in one place and furiously veral years before observing its consequen- stomp the lake botton at a rate of more ces for P. tricolor (Table 4). Typically, than 200 "stomps" per minute. The beha- P. chilensis walks steadily along, P. andinus vioral differences among the species are meanders, stopping occasionally, and P. correlated with the escape ability of their jamesi walks very slowly, often spending respective prey (Hurlbert, López and many seconds working the sediments Stein, unpubl. data).

TABLE 3 Association of Wilson's phalarope with flamingos on Laguna Chulluncani. 22 February 1979 La asociaci6n de P. tricolor con los flamencos en Laguna Chulluncani. 22 febrero 1979

Number of phalaropes in attendance Flamingo Activity per flamingo (k) species 2 3 4 5 6 7

Number of flamingos attended Total number of by k phalaropes flamingos observed

Morning hours) jamesi feeding andinus feeding 23 23 chilensis feeding 5 36 14 7 2 1 1 66 ch ilensis resting

Afternoon hours) chilensis feeding 128 41 6 2 177

TABLE4 Rates of walking of the three flamingo species during continuous feeding Velocidad de caminar de las tres especies de flamenco durante alimentaci6n continua

Walking rate (steps/minute) date no. birds. standard and species mean observed deviation range (x) (n) (s)

de Chile, 25 Nov. 1975 (26º22' 69º12' W) jamesi 8.7 7 3.9 4-16 andinus 29.2 6 13.4 chilensis 43.1 7 14.6 23-68

Salar. de Aguas Calientes III, Chile. 29 Nov. 1975 (25º00' 68º38' W) P. jamesi 8.5 4 4.7 5-15 andinus 18.4 5.6 chilensis 37.3 9 23-55

Laguna Chulluncani, Bolivia, 12 Dec. 1976 jamesi 7.2 2-26 P. andinus 25.8 9 9.0 12-36 chilensis 54.3 7 4.7 WILSON'S PHALAROPE IN THE CENTRAL ANDES 55

DISCUSSION suggested by the earlier-mentioned obser- vations at Mar Chiquita (C. Olrog, pers. Abundance. The salt lakes of southern comm.), lowland Argentina serves as win- Bolivia clearly constitute one of the major, tering grounds for a large fraction of this if not the major, wintering grounds for species. It is likely that lowland Argentina P. tricolor. It also seems likely that during served as wintering ground for greater the austral summer this species is the most numbers of P. tricolor in the past than abundant bird in the entire Andean alti- at present. Expansion of the human popu- plano. Yet there are no previous published lation, agriculture, fish introductions and records of its presence in the Bolivian other habitat-modifying activities of man altiplano and very few for elsewhere in probably have destroyed the suitability the Central Andes -Cumbres Calchaquies, of many areas for P. tricolor, just as it has to the east of Tucuman, Argentina (Dab- rendered them unsuitable as nesting areas bene 1920). Lago Titicaca and a few for P. chilensis (Johnson 1965 ). Of course other localities in Peru (Morrison 1939, many other birds are likely to have been Holmes 1939), and Lago (Embalse de) affected similarly by such change. Caritaya, Chile (Johnson 1965). Our suggestion that it may be the most Salinity. Because P. tricolor overwinters abundant bird in the altiplano is based mostly in inland areas, it has been assumed on the size of flocks we observed and our to be less associated with saline waters extensive inspection, flamingo censusing, than are the other two phalaropes, which and invertebrate sampling of Central are oceanic except during the breeding Andean lakes between the latitudes of season. Murphy ( 1936) stated that P. Cuzco, Peru (14° S) and Copiapo, Chile tricolor "is found inland to a greater ex- (27° S) (Hurlbert 1978, 1981, unpubl. tent than on salt water", failing to recogni- data; Hurlbert and Keith 1979). Using ze that the saltiest waters in the world abundance of aquatic invertebrates and are inland lakes. Palmer ( 1967) refers P. chilensis as a criterion, we predict that to it as "A fresh-water bird, with rudi- large numbers (> 1000 individuals) of mentary salt gland (rather than large P. tricolor should sometimes be found in gland as in the marine phalaropes)". the following additional lakes: in Bolivia, Our data (Table I) show that on its winter- at Laguna Khara (Fig. 1), Lago Soledad ing grounds in Bolivia P. tricolor confronts (17°44' S, 67°22' W), Lago Uru-Uru, much higher salinities than do the ocean- Lago Poopo; in Chile, at Salar de Pujsa going phalaropes (ocean salinity is about and Salar de Aguas Calientes I and II (Fig. 35/oo). Moreover, at the beginning of its 1); and in Peru, at Lago Parinacochas southward migration P. tricolor assembles (15° 17' S, 73° 42' W) and Lago Salinas in large numbers and feeds at a few large, (14° 59' S, 70°07' W). Other waterbodies highly saline lakes in northern United not visited by us, such as Laguna Vilama, States (J. J ehl, pers. comm.): Mono Lake, Argentina (Fig. 1) and the temporary California (80-90°/oo), Great. Salt Lake, rainwater pools of the central and northern Utah (110-270/oo), and Lake Abert, altiplano possibly serve as winter feeding Oregon (35-105/oo). In these respects, it is grounds, too. one of the most "salt water" birds in the The second most abundant bird in the world. However, at all of these lakes, fresh- Andean altiplano probably is P. chilensis, water springs, seeps and/or streams are of which there are estimated to be about present at the lake margins and so the 350,000 individuals (Kahl 1980). However birds never need to go for long periods a significant but unknown fraction of without drinking freshwater. That undoub- these occurs in lowland areas of Chile and tedly accounts for the small size of their Argentina. salt gland, which, in any case, is capable Though there are numerous published of rapid enlargement (J. J ehl, pers. comm.). records of P. tricolor for lowland areas of Argentina (summarized in J. J ehl, in Interaction with P. chilensis. The advantage manuscript), none of them report large to P. tricolor of consorting with P. chilensis numbers. Nevertheless we believe this is clear. P. tricolor in some situations will may reflect only insufficient inspection submerge head and neck while feeding of the most suitable regions and lakes and (Bent, 1927), but even where water is only that, as has been long suspected and as I 0 em deep, they cannot feed on inverte- 56 HURLBERT ET AL.

brates that remain at the bottom of the their distribution and movement among water column or on the sediment surface. lakes in the region, and their physiological The vigorous movements of feeding P. condition (and hence chances for survival) chilensis bring invertebrates within the at the start of their migration back to smaller bird's reach. North America. Whether man, by intensi- P. tricolor has been observed to cluster vely harvesting the eggs of P. chilensis about feeding American avocets (Recur- and thereby reducing its numbers, has virostra americana Gmelin) in exactly indirectly increased the abundance of the manner they do with P. chilensis (Wi- aquatic invertebrates in altiplano lakes lliams 1953). Perhaps they sometines and thus made them more favorable as associate with the Andean avocet (Recur- wintering grounds for P. tricolor is open virostra andina Philippi & Landbeck) to speculation. It is clear, however, that several dozen of which are often present. intensive harvesting is occurring (Hurl- on individual altiplano lakes, but we have bert and Keith, unpubl. data) and it prob- not observed this. The ocean-going phala- ably has been going on since man arrived ropes, Phalaropus lobatus (Linnaeus) and in South America. The small colony P. fulicarius (Linnaeus), occasionally dine observed at Laguna Herrera in 1978·79 in the commotion of even more distant was the only P. chilensis colony (of the relatives, feeding on invertebrates chased several we have observed in the altiplano) to the surface by schools of fish or carried that escaped the attention of the egg there by the turbulent wake of swimming harvesters and produced flamingo chicks. whales (Bent 1927). However, we do not have direct evidence P. chilensis is very tolerant of P. trico- that egg harvesting has reduced the num- lor despite the insolence with which the bers of P. chilensis or even that P. chilensis latter dodges about the flamingo's legs. has declined in abundance. We have never observed any hostile re- action of the flamingos to the phalaropes. Yet when present in large numbers P. ACKNOWLEDGMENTS tricolor must markedly reduce food avai- Many individuals have assisted this work. S. Pezzani and lability to P. chilensis, and vice versa. L. Robres accompanied us on our first trips to Salar de They are clearly competitors. Carcote. L. Peña introduced us to the lakes of the Boli- Even by itself P. chilensis can rapidly vian puna, and T. Saire guided us to the most spectac- ular P. tricolor flocks. A. Barrero, fils et père, provided deplete the invertebrate populations of a invaluable logistical support and the hospitality of Mina shallow lake (Rooth 1965 ). At Laguna Laguna Verde. G. Bejarano and the Bolivian Ministry 2 Herrera (area = I km ), Bolivia on Decem- of Agriculture facilitated our studies in many ways. ber 25, 197 6 the density of copepods We are grateful to all these persons and to many other (Boecke/la poopoensis copepodids) was inhabitants and authorities in the region for their help. We thank F. Reiss and I.A.E. Bayly for identification 382/liter. Fourty-three days later (Fe- of chironomids and copepods, respectively. This paper bruary 6, 1979) the copepod density was has been improved by comments and additional data II /liter. Nauplius larvae declined corre- provided by B. Araya, J. Fjeldsa, J. Jehl, M.P. Kahl, spondingly. The evidence is only circum- C. Olrog, and M. Sallaberry. This work was supported by grants from the National stantial, but it seems highly probable that Geographic Society and the U.S. National Science Found- this reduction was the result of predation ation (DEB 76-02888). by the approximately 150 pairs of P. chilensis that were nesting at this lake. Some P. tricolor (numbers unrecorded) LITERATURE CITED were feeding in close association with BENT AC (1927) Life histories of North American the P. chilensis on February 6, and may Shorebirds, Part 1. U. S. National Museum, have contributed to the copepod decline. Bulletin 142: 1-359. There is evidence that predation by the DABBENE R (1920) Notas sobre los chorlos de Norte Red Phalarope may have America que inviernan en Ia Republica Argen- (P. fulicarius) tina. El Hornero 2: 99-128. a significant impact on the zooplankton DODSON SI & DL EGGER (1980) Selective feeding of of shallow ponds in the Arctic tundra Red Phalaropes on zooplankton of Arctic (Dobson and Egger 1980). ponds. Ecology 61: 755-763. By its effects on prey abundance, P. FJELDSA J (1981) A comparison of bird communities in temperate and subarctic in northern chilensis may influence P. tricolor in se- Europe and the Andes. Proceedings of the veral ways. It could affect the number Second Nordic Congress of Ornithology, 1979: of P. tricolor that overwinter in the region, 101-108. Norwegian Ornithological Society. WILSON'S PHALAROPE IN THE CENTRAL ANDES 57

HOLMES PF (1939) Some oceanic records and notes on KAHL MP (1980) Population ecology of the flamingos of the winter distribution of phalaropes. Ibis 81: the world. National Geographic Society Research 329-342. Reports 12: 407-415. HURLBERT SH (1978) Results of five flamingo censuses MEINERTZHAGEN DSO (1925) The distribution of the conducted between November 1975 and De- phalaropes. Ibis 1925: 325-345. cember 1977. Andean Lake and Flamingo MORRISON A (l.939) The birds of the Department of Investigations, Technical Report No. 1, 16 pp. Huancavelica, Peru. Ibis 3: 453-486. San Diego State University, San Diego, Cali- MURPHY RC (1936) Oceanic birds of South America, fornia. Vol. II. American Museum of Natural History, HURLBERT SH (1981) Resultsofthreeflamingocensuses New York. conducted between December 1978 and July MYERS JP & LP MYERS (1979) Shorebirds of coastal 1980. Andean Lake and Flamingo Investiga- Buenos Aires Province, Argentina. Ibis 121: tions, Technical Report No. 2, 9 pp. San Diego 186-199. State University, San Diego, California. OLROG CC (1967) Observaciones sabre aves migratorias HURLBERT SH (1982) Limnological studies of flamingo del hemisferio norte. El Hornero 10: 292-298. diets and distributions. National Geographic PALMER RS (1967) Species accounts. In: Stout, GD Research Reports 14: 351-356. HURLBERT SH & CCY CHANG (1983) Ornitholimno- ed., The shorebirds of North America. Viking Press, New York. 143-267. logy: effects of grazing by the Andean flamingo (Phoenicoparrus andinus). Proceedings of the ROOTH J (1965) The flamingos on Bonaire. Publications U. S. National Academy of Sciences 80:4766- of the Foundation for Scientific Research in 4769. Surinam and the Netherlands Antilles, No. 41, HURLBERT SH & JO KEITH (1979) Distribution and 151 pp. Utrecht, Netherlands. spatial patterning of flamingos in the Andean WETMORE A (1926) Observations on the birds of Argen- altiplano. Auk 96; 328-342. tina, Paraguay, Uruguay and Chile. U.S. Natio- JOHNSON A W (1965) The birds of Chile and adjacent nal Museum, Bulletin 133: 1-448. regions of Argentina, Bolivia and Peru. 2 vols. W1LLIAMS GG (1953) Wilson phalaropes as commensals. Platt Establecimientos Gnificos, Buenos Aires. Condor 53: 158.