SHORT COMMUNICATIONS 145 includingone inside a nestcavity. Six other islandsshowed penetrated in the present study, however, suggeststhat no indication of abnormally low nestingsuccess, and were ermine are quite capable of reaching coastal islands in probably not visited by ermine. years when their populations are high, and that such in- The major effects of ermine on guillemot breeding ap- vasionsmay be more frequent than the paucity of records pearedto be discouragementof egg-layingand a reduction suggests. of hatching successdue to nest abandonment. Many cav- ities on Black, Yellow, and Green islands which were oc- I thank W. Cairns, V. Friesen and H. Kaiser for field cupied in previous years were not used in 1983. Eggslaid assistance,and A. J. Gaston, M. B. Fenton and P. Ewins on islands where ermine presencewas confirmed or sug- for commenting on the manuscript. This study was sup- gestedwere often displacedfrom the nest cup and lacked ported by the Canadian Department of Supply and Ser- the shiny appearancewhich is normal for regularly incu- vices and the Canadian Wildlife Service. bated eggs(pers. observ.). Guillemots may have been re- LITERATURE CITED luctant to enter their nests if they had seen an ermine in the colony, which would explain the reductionin both egg- BANFIELD,A. W. F. 1974. The mammals of Canada. laying and hatching. Univ. of Toronto Press. Ermine visited islands as far as 1.6 km from the main- BIANIU,V. V. 1967. Gulls, shorebirdsand alcids of Kan- land (Pitsulak City), and may have reached Kingituayu dalaksha Bay. Israel Program for Scientific Transla- Island as well (2.2 km). This latter island is the highest of tions, Jerusalem. the 12 examined (22 m above sea level), which suggests FINERTY,J. P. 1980. The population ecology of cycles that conspicuousnessfrom the mainland may increase in small mammals. Yale Univ. Press, New Haven, likelihood of visitation. CT. Ermine must have reachedthese islandseither by cross- GASTON, A. J., D. K. CAIRNS,R. D. ELLIOT,AND D. G. ing the ice or by swimming. If they came over the ice, they NOBLE. In press. A natural history of DiggesSound. must have arrived at least two weeks before laying, since Can. Wildl. Serv. Rep. Ser. heavy ice disappearedfrom the area in early June. If er- KARTASCHEW,N. N. 1960. Die Alkenvogel des Nordat- mine were on islands during the pre-laying period, they lantiks. Die Neue Brehm-Btlcherei 257. Wittenberg, might have subsistedon eggsabandoned the previousyear. West Germany. Such food, however, would not have been available on LACK, D. 1967. Interrelationships in breeding adapta- PitsulakCity, becauseall unhatchedeggs were clearedfrom tions as shown by marine birds. Proc. XIV Int. Or- the island the previous fall. Ermine may instead have nithol. Congr. (1966):3-42. reached the islands by swimming. A local Inuit told me SETON,E. T. 1929. Lives of game animals. Doubleday of seeing an ermine swimming in salt water, and Seton Doran, New York. (1929) reported anecdotalaccounts of the species’ ability STORER,R. W. 1952. A comparison of variation, be- to swim. havior and evolution in the sea bird genera Uria and Ermine numbers are known to vary with those of lem- Cepphus.Univ. Calif. Publ. Zool. 52:121-222. mings (Finerty 1980), and the 1983 boom in ermine pop- ulations that led to the invasion of coastal islands in the BiologyDepartment, Carleton University, Ottawa KlS 5B6, study area may have been related to the lemming cycle. Canada. Presentaddress: Newfoundland Institute for Cold Although ermine and guillemots(Cepphus spp.) are broad- OceanScience, % BiologyDepartment, Memorial Univer- ly sympatric (Storer 1952, Banfield 1974), reports of pre- sity OfNewfoundland,St. Johns,’ NewfoundlandAlB 3X9, dation by small weaselson guillemots are rare (e.g., Kar- Canada. Received 7 December 1983. Final acceptance12 taschew 1960, Bianki 1967). The number of colonies September 1984.

The Condor 87:145-147 0 The Cooper Ornithological Society 1985 and derive a major portion of their diet as ectoparasites taken from the skins of these animals (Ma&worth-Praed and Grant 1955). OBSERVATIONS OF SCRUB JAYS In North America, examples of such protocooperation CLEANING ECTOPARASITES are rare. Although Cattle Egrets(Bubulcus ibis). cowbirds (Molothrus spp.); and occasionallyother blackbirds often FROM BLACK-TAILED DEER perch on the backs of cattle or other large grazers, they are thought to use the cattle only as “beaters” to stir up insectfood from the grass(Heatwole 1965,Dinsmore 1973). FRANK R. ISENHART It is not clear that cowbirds or Cattle Egrets ever pick

AND insectsoff the animals themselves.This behavior is, there- fore, better consideredan example of commensalismrath- DAVID F. DESANTE er than protocooperation. In contrast, Black-billed Mag- pies (Pica pica) pick and eat ticks from the backs of elk (Cervurcanadensis) and mule deer (Odocoileushemionus), Instances of protocooperation between birds and other an example of true protocooperation(Linsdale 1946). large animals are relatively uncommon, although certain Here we report a seriesof observationsof an apparently examplesare quite well-known. These include the Greater protocooperative associationbetween Scrub Jays (Aphe- Honeyguide(Indicator indicator)leading the honey badger locoma coerulescens)and Columbian black-tailed deer (Mellivora capensis)to a beehive (Skead 195l), the Egyp- (Odocoileushemionus columbianus) in which jays regu- tian Plover (Pluvianusaegyptius) picking leeches and small larly picked ticks and perhapsother ectoparasitesfrom the bits of food from the open mouths of crocodiles (Howell skin of deer. Dixon (1944) first reported such behavior 1979), and the Small Ground Finch (Geospizafuliginosa) between a Scrub Jay and a California mule deer (0. h. grooming ticks from the skin of the marine iguana (Am- cahfornicus)in SequoiaNational Park, California. Schulz blyrynchuscristatus; Amadon 1967). Uniquely among and Budwiser (1970) reported another Scrub Jay presum- birds, the two African speciesof oxpeckers(Buphagus spp.) ably taking ectoparasitesfrom the back of a black-tailed habitually associatewith herds of large grazing animals deer 4 km northeast of Alpine Lake in Marin County, 146 SHORT COMMUNICATIONS

California. For both cases, single observations were re- deer entered the raised open area; one bird immediately ported of individual ScrubJays alighting on a deer’s back, flew onto its back, while the other remained on its perch. picking at the deer’s skin, and flying off. We have tallied The deer remained motionless even though it faced an eight additional cases,five of them involving singlejays observer, 150 m away. Cleaning followed in the sequence and three involving pairs of jays. These observationsin- describedabove, after which the jay returned to its partner clude the interaction of at least five different Scrub Jays and the deer departed. After 3 min, two additional deer (two of which were individually color-banded)and 17 deer, appearedon the outcroppingand were cleanedin the same althoughwe do not know exactly how many different deer manner as the first. We do not know whether or not the were involved becausethey were not marked. jay that performed this cleaning was the same individual The first observation was made in spring 1982 by Jack that cleaned the first solitary deer. When this jay was fin- Swenson on the undisturbed, mature part of the 36-ha ished with both deer, it returned to its partner and both coastalscrub study plot of the Palomarin Field Station of deer re-entered the thick vegetation beside the tree. This the Point Reyes Bird Observatory(PRBO), at the southern interaction lasted 20 min and occurred in full view of a end of the Point Reyes National Seashore(PRNS), Marin human observer. The observer’s presencedid not seem to County, California. The remainingobservations were made affect the deer’s or the jays’ behavior, as we confirmed between January and May 1983-by Isenhart, David Sie- subsequentlyby using more inconspicuousobservation mens, and David Forma. Five of theseincidents occurred points and even a blind, the same behavior was seen as near the location of the 1982 sightingand three occurred before. about 2 km northwest on a recently burned coastal scrub Corvids are widely accepted as highly intelligent birds study plot. Both areas were situated on variously gentle (Linsdale 1937) with well-developed learning capabilities to steep south-southwest-facingslopes that are immedi- (Brower et al. 1970). At Palomarin, young color-banded ately adjacent to the Pacific Ocean. The vegetation there ScrubJays remain with their parentsfor up to five months consistedprimarily of extensiveareas of successionalstage after fledgingand could well learn to clean deer during this coastal scrub (recovering from grazing and limited culti- time. Young Scrub Jays disperseduring their first fall or vation prior to 1965), interspersedwith smaller patches winter and may go surprisingly far. One Scrub Jay that of mature undisturbed coastal scrub. Coyote bush (Bac- was color-banded at Palomarin during its first winter was charispilularis, var. consanguinea)was the predominant recapturedfour years later as a resident of Carson Ridge, shrubin both areas,but substantialnumbers of other coastal 11 km east-northeastof Palomarin. This location is ap- scrub speciesalso occurred, especiallyin the undisturbed proximately 3 km northeast of Alpine Lake. and must be areas.One plant associate,blueblossom (Ceunothus thry- closeto the location where, in 1965, a ScrubJay was seen siflorus),is known to be an especiallyfavored waiting place atop a deer bv Schulz and Budwiser (1970). Thus. it is for the Pacific Coast tick (Dermacentor occidentalis),a possible that young jays could learn ‘cleaning behavior common parasite of both humans and deer at Palomarin from their parents and spread it among other jays when (Robert S. Lane, pers. comm.). The open coastalhills are they disperse from their natal territories. Several other borderedto the eastand north by douglas-fir(Pseudotsuga pairs of color-banded Scrub Jays at Palomarin, however, menziesir) forest and, in the arroyos and canyons, by occupy territories adjacent to the territory of a “learned” broadleaf evergreenwoodland. The closejuxtaposition of pair of jays, and have never been seen to alight upon a these various habitats, and the protection afforded by the deer, although deer are equally abundant and accessible PRNS, combine to support large resident populations of in their territories. Certain of these jays are now eleven both Scrub Jays and black-tailed deer. yearsold and have been intensively studied for more than In a typical interaction between a jay and a deer, the five years without any evidence of protocooperationwith jay would land on the neck, back, or head of the deer and deer. To some degree,the deer’s role in this interrelation- proceedto “clean” the entire dorsal surfaceof the animal, ship may also localize the behavior. Linsdale (1946) noted includingthe upper flanks and hind limbs, tail, back, neck, that some deer would not cooperate when magpies at- head, and antlers. Every few seconds,the jay would peck tempted to clean them and would push the magpies off at the deer’s skin. Two types of peckswere used. One type their backswith their noses.Perhaps some deer are more was delivered forcefully and the other softly, usually pre- receptive to this form of protocooperationthan others. cededby a brief waiting period. These different pecksmay During one 2-h observation, over 80% of a singlejay ’s have reflectedtwo types of food selection;hard pecksmay time was spent looking for or interacting with deer. When have been for ticks, and soft ones for keds (hippoboscid two deer and two jays were together, only one jay cleaned flies) and deer flies (Tabanidae). Throughout this period, while the other watched from a nearby tree. Presumably, the deer would remain motionless. In one case, the deer the cleaning behavior must be highly beneficial for such was moving when the jay alighted on its back, but stopped an investment of time and energy to be worthwhile. Fur- and stood on three legs during the 5-min cleaningperiod. thermore, this behavior has been observed at Palomarin On three ocasions,the deer extended their ears while the only from late winter through spring, when ticks are at jays concentrated on that area of the head. Presumably, least ten times more numerous (on humans) than during this allowed the birds to exploit an otherwise inaccessible the rest of the year. We do not know to what extent this area of the deer’s ear and head. In all cases,the jays called protocooperationbetween ScrubJays and black-taileddeer before and after, but never during, the interaction. exists throughout the year or exists in other coastal or In a typical interaction between a pair ofjays and a deer, inland populations. the birds perched atop a small tree (4-8 m high) or flew The observations reported here were made as part of from treetop to treetop. In either case, the jays usually PRBO’s ongoing Coastal Scrub Avian Ecoloav Proaram. called loudly. When a deer appeared or was found, one -. _ , which is financially supported primarily by the member- jay alighted on the deer’s back and behaved as described ship of the Point Reyes Bird Observatory. Robert S. Lane, above, while the other remained perched in the top of a Specialist,Division of Entomologyand Parasitology,Uni- nearby tree (5-8 m). When the one jay completed its clean- versity of California, Berkeley,kindly identified specimens ing, it returned to its partner (the known mate in the case of the tick. Helpful comments on this manuscript were ofthe color-bandedpair ofjays), and the pair either moved made by T. Darrow, D. McKenzie, and F. Pitelka. This on to another deer or waited for one to appear. is Contribution No. 269, PRBO. Activity suggesting“cleaning station” behavior was also noted with pairs of jays. In one case, two jays were seen LITERATURE CITED perchedatop a douglas-fir.Thick coastalscrub surrounded the tree except where a rocky outcropping (8 x 4 m) ex- AMADON,D. 1967. Galapagosfinches grooming marine tended out from the base. While the jays were calling, a iguanas.Condor 6913Il. SHORT COMMUNICATIONS 147

BROWER,L. P., B. S. ALPERT,AND S. C. GLAZIER. 1970. sonia), p. 133-154. In A. C. Bent, Life histories of Observational learning in feeding behavior of Blue North American jays, crows, and titmice. U.S. Natl. Jays (Cyanocittacristata). Am. Zool. 10:475-476. Mus. Bull. 191. (Dover reprint, 1964). DINSMORE,J. J. 1973. Foraging successof Cattle Egrets, MACKWORTH-PRAED, C. W., AND C. H. B. GRANT. 1955. Bubulcusibis. Am. Midl. Nat. 891242-246. Birds of eastern and northeastern Africa. Vol. II. DIXON, J. S. 1944. California jay picks ticks from mule Longmans Green and Co., London. deer. Condor 46:204. SCHULZ, T. A., AND P. D. BUDWISER. 1970. Scrub Jay HEATWOLE,H. 1965. Some aspectsof the associationof possiblyfeeding on ectoparasitesof a black-taileddeer. Cattle Egretswith cattle. Anim. Behav. 20:42 l-424. Calif. Dept. Fish Game Bull. 56 (January). HOWELL,T. R. 1979. Breeding biology of the Egyptian SKEAD, C. J. 1951. Notes on honeyguidesin southeastern Plover, Pluvianusaegyptius. Univ. Calif. Publ. Zool. Cape province, South Africa. Auk 68:52-62. 113. LINSDALE,J. M. 1937. Natural history ofMagpies. Pacific Point ReyesBird Observatory,4990 ShorelineHighway, Coast Avifauna 25. Stinson Beach, California 94970. Received 13 January LINSDALE,J. M. 1946. American Magpie (Picapica hud- 1984. Final acceptance7 November 1984.

The Condor87:147-148 24 l-242, 1971) describeda similar instanceof feeding by 0 The CooperOrnithological Society 1985 Pied-billed Grebes and Snowy Egrets (Egretta thula) in which the grebeswere chasingsmall fish into shallowwater A MUTUALISTIC FEEDING where they were capturedby both species,the grebesben- efitting by the herons “chasing” fish from their refuge in ASSOCIATION BETWEEN vegetation. Mueller et al. (Auk 89:190, 1972) reported BOAT-TAILED GRACKLES similar interactions between Pied-billed Grebes, a Tri- AND PIED-BILLED GREBES colored Heron (E. tricolor),and a Snowy Egret. Paulson (Auk 86:759, 1969) reviewed examples of feeding asso- ciations between other grebe species and other aquatic JEROME A. JACKSON birds. The observation reported here is novel in that it involved interaction with a terrestrial bird species, and unusualprey for the grebes. Similar involvement of multiple individuals of two bird At noon on 25 December 1978, I observed a feeding as- specieswas described by Clark (Fla. Field Nat. 6:45-46, sociation between Pied-billed Grebes (Podilymbuspodi- 1978) for a feeding associationof American White Peli- ceps)and female Boat-tailed Grackles (Quiscalusmajor) on Horn Island, approximately 10 km south of Ocean cans (Pelecanuserythrorhynchos) and Wood Storks (Myc- teria americana), and by Rodgers (Fla. Field Nat. 6:44- Springs, Jackson County, Mississippi. My attention was drawn to the birds becauseof their numbers, proximity 45, 1978) for Brown Pelicans(Pelecanus occidentalis) and Wood Storks.Both ofthesecases also recognized one species to one another, and their frenzied behavior. Observations were made with a 20 x spotting scopefrom 30 m away at (the Wood Stork) as the primary beneficiary of the asso- ciation. Those associationsmight also have been mutu- a ca. l-ha freshwaterpond fringed with yaupon (Ilex vom- alistic, however, becausefish that escapedthe Wood Storks itoria) and low (~0.3 m) grassesand herbaceousvegeta- tion. When first seen,two grebeswere feeding within 1 m were probably often herded back into the path of the pel- of one another and within 0.3 m of a grassy shoreline. ican assemblages. Suchinteractions involving multiple individuals of each Water depth did not exceed 0.2 m and the grebes were predatorspecies herding multiple prey individualsare likely capturing prey from the surface. A tight group of eight to be mutualistic. Various authors have referred to such female Boat-tailed Grackles was clustered at the water’s associationsas “cooperative” (e.g., Leek 1971) or “com- edgenear (often within 0.1 m) the grebes.The mixed group mensal” (e.g., Paulson 1969, Clark 1978, Rodgers 1978). moved steadily along the shore at about 1 m/10 s. None of the interspecificassociations there describedsug- As the grackles worked their way through the grass, gested that the relationship resulted from active associa- climbing, hopping, and flying, I sawnumerous grasshopper tion by both species.Rather it appearsthat one, the grebe nymphs jumping in front of them. Some were captured or the Wood Stork in the casesdescribed above, was at- by the grackles,others escapedto the water where many tracted to the feeding activities of the other. Thus “co- were caught by the grebes, and still others escapedback operative” seemsan inappropriate descriptor.In all of the to land either to be eaten by waiting grackles or to be casesdescribed here, however, both speciesprobably ben- chasedback to the water. Becauseof the rapid movement, efitted from the activity, albeit the grebesand storks per- the number of birds, and the partially obscuringvegeta- haps more so. Thus, the associationsare more than “com- tion, it was difficult to determine capture rates. For brief mensal.” It seemssignificant that in none of the caseswas periods when I was able to keep one bird in view, grackles interspecificevasive or aggressiveactions observed. This caught grasshoppersat an average rate of one every 9 s (range 5-17 s, n = 11) and grebescaught grasshoppersat supportsthe notion that the associationswere mutualistic. a rate of one every 11 s (range = 5-23 s, n = 14). These It is easy to imagine, however, that under circumstances of more distant or closer associationand/or decreasedor peak rates were for individuals closestto the water’s edge, increased numbers of grebes, the relationship might be- capture rates seemed slower for birds more distant. This come one of commensalism or kleptoparasitism, respec- feeding frenzy continued for nearly 20 min, at which time tively. Interspecificfeeding associations thus seemto form the group arrived at a densecattail (Typha latifolia) stand. a gradedseries from commensalismto mutualism to klep- The gracklesthen flew off as a group and the grebesdis- toparasitism, depending on the closenessof the birds and appearedinto the cattails. the numbers of the “benefitted” species. Although the grebeswere the primary beneficiaryof the feedingassociation, the observedbehavior was mutualistic I acknowledgeboth logistic and financial support from rather than merely commensal. Most captures by both the U.S. National Park Service,GulfIslands National Sea- speciesresulted from flushingof the insectsback and forth shore, for my studies on the National Seashoreislands. between the land and the water. Leek (Am. Midl. Nat. 86: Stephen Cofer-Shabica, Jim Ray, and numerous park