SHORT COMMUNICATIONS 559

available to wrens for shelter or warmth. I did not record Washington’s history occurred in January, 1950. On Ta- air temperatures,but three consecutivenights were clear toosh, the maximum recorded temperature between 13- with strong (10-25 knot) winds from the east. Air tem- 18 January was 28”F, and the minimum was 14°F (U.S. peraturesat the nearestNational Weather Service station Dep. of Commerce 1950). Wrens were abundant in 1956 (Quillayute; approximately 55 km to the southeastand at (F. Richardson,pets. comm.). From 16-l 8 December 1964, an elevation comparable to Tatoosh) were recorded as air temperatureson Tatoosh were generallybelow freezing, (date; max./min. degreesF): 28 December, 32/16; 29 De- ranging from 33-14°F (U.S. Dep: of Commerce 1964).-I cember, 29/19: 30 December. 26/20: 3 1 December. 28/9: heard wrens in June. 1968. I know of no records of dis- 1 January, 32jlO (U.S. Dep.‘of Commerce 1978, 1979): appearanceof Winter Wrens from mainland sites during On my trips to the island on 16-21 March and 26-29 any of these intervals. April 1979, I found neither wrens nor the other two con- Extreme climatic events, suchas freezes,are recognized spicuous,resident passerines,Song (Melospiza melodia) as important determinants of island avifaunas. Winter and Fox (Passerella iliaca) sparrows. Wrens can be abundant, permanent, and conspicuousres- My observationson subsequenttrips suggestedthe fol- idents on islandslike Tatoosh. In the last decade,however, lowing recolonizationpattern. No wrenswere seenor heard they have been absent as breeders50% of the time (1979- in 1979 (n = 12 visits), 1980 (n = 9), 1982 (n = lo), or 1984)and, since 1950, I estimate that breedingwrens have 1983 (n = 12). In 1981, wrenswere heardinFebruaryand been absent a maximum of about 40% of the years. March, but were neither seennor heard on the subsequent I am grateful to the many ornithologistswilling to share sevenvisits. In1 984, wrens were heard in mid-March, and their field notes with me, to the U.S. Coast Guard and the on nine of 10 visits between then and mid-November. Makah Tribal Council for permission to do research on The density of singingmales (five) was roughly about 50% Tatoosh, and to the National ScienceFoundation for sup- of the previous maximum. Wrens have been conspicuous port. on all (six) 1985 trips to date. The other passerinesof interest were active at a feeder that was established in LITERATURE CITED October, 1979, and had returned to pre-freeze abundance by late June, 1980 (G. B. van Vliet, pers. comm.). ARMSTRONG,E. A. 1955. The Wren. Collins, London. These observationsare compatible with what is known BATTEN,L. A. 1980. Some recentbird populationchanges of Winter Wren biology. Bent (1948, and referencesthere- in Britain. Ibis 122:420. in) recordeda number of invasions or extinctions on hab- BENT,A. C. 1948. Life historiesof North American nut- itable islands in Alaska. In Europe, the same speciesoc- hatches,wrens, thrashersand their allies. U.S. Natl. cupiessimilar vegetation-densehabitats (Armstrong 1955). Mus. Bull. 195. Although few data exist on fluctuations in insular popu- JEWETT,S. A., W. P. TAYLOR,W. T. SHAW,AND J. W. lations, on the mainland, wrens seem especially suscep- ALDRICH. 1953. Birds of Washington State. Univ. tible to cold weather, and show precipitous population of Washington Press, Seattle. declines during severe winters (Batten 1980, Williamson U.S.DEPARTMENTOFCOMMERCE.1950,1964,1978,1979. 1981). On Tatoosh, the 6-year interval noted between ex- Climatological Data, Washington. Weather Bureau. tinction, presumablydue to some combination of cold and WILLIAMSON,M. 1981. Island populations. Oxford. starvation, and recolonization probably indicates the dif- ficulties experienced by a relatively sedentary speciesin Department of Zoology, University of Washington, Seat- invading nearby habitable terrain that is isolated by a tle, Washington 98195. Received 29 December 1984. Fi- water gap. For instance, the coldest weather in western nal acceptance8 July 1985.

The Condor87559-561 Individuals in the study population have been color- 0 The Cooper Ornithological Society 1985 banded since 1972. The nesting attempt describedherein was begun on 13 May 1983 by a four-year-old male and BEHAVIOR AT A PINYON JAY a three-year-old female in their second breeding season NEST IN RESPONSE TO together. The nest was placed 8.5 m high in a 10-m pon- derosa pine tree (Pinus ponderosa). Five eggswere laid, PREDATION all of which hatched on 30 May. My results reported here were based on 20.6 h of nest observations over six days (10.6 h before predation and JOHN M. MARZLUFF 10.0 h after predation). I observedthe nest from a distance of 15 m in a canvas blind. Before predation, I watched the nest for 4.6 h on the morning of 13 June, for 4.0 h Predation is a major causeof nest failure in a population after noon on 14 June, and for 2.0 h before noon on 15 of Pinyon Jays(Gymnorhinus cyanocephalus) that inhabit June. Following predation, I watched the nest for 2.0 h Flagstaff,Arizona (Marzluff 1983). Jaysin this population before noon on 15 June, for 3.0 h after noon on 15 June, reduce predation by nesting colonially, concealing their and for 5 h before noon on 16 June. nests (Marzluff 1983), and grouping fledged young in a On 15 June,both parentsarrived in the nest tree together crechethat is guardedby sentinels(Balda and Balda 1978). at 08:24, and proceededto feed and clean the 16-day-old Typically, they mob predatorsand deserttheir nests(Clark young. They departed 2 min later, and an adult American and Gabaldon 1979) after acts of predation. I report here Crow appeared at the nest at 08:58. All nestlingsbegged the responseof one pair of jays to the partial removal of from the crow, as they would from their parents (see their brood by an American Crow (Corvus brachyrhyn- McArthur 1982 for a description of begging).The crow chos). The predationevent spanneda two-day period,which jabbed one young bird repeatedly in the head and neck enabled me to contrast the parents’ behavior during the region until it was dead. During this time, all nestlings event with their behavior before the event. I also compare called harshlv. The crow then araspedthe dead youngwith their behavior to the behavior of other jays that were not its bill and carried it from thenest. The crow’s visit iasted detected by predators. about three minutes. 560 SHORT COMMUNICATIONS

While the parents were still gone, the crow returned to strikingit on the back with feet and bills. The adultsuttered the nest at 09:22 and killed another nestling. A Steller’s begging-likequays (Balda and Bateman 1973) in addition Jay (Cyanocitta stelleri) discovered the crow there and, to normal mobbing racks and rackas. The nestlings vo- with crest erect, approached to within 0.3 m of it while calized continuously. The parents forced the crow from utteringthe shookscolding call (Brown 1964). The Steller’s the nest for 90 s. It then returned to the nest, graspedone Jay called approximately 90 s before the crow departed young alive in its bill, flew to the ground 10 m from the with the dead young. Both Pinyon Jay parents returned nest, and killed it. During this time, the crow was unre- as the crow was leaving, mobbed it, and chased it out of sponsiveto continued mobbing and dive-bombing by the sight. I heard Pinyon Jaysscolding in the area for the next parents. It left the area at 11:30, 3.5 min after taking the 4 1 min. Afterward, both parentsreturned to the nest where young jay. The remaining two young departed the nest they fed and cleanedthe remaining three nestlings.I heard prematurely (age 17 days, instead of 21 days) during the scolding in the area for 10 min following this nest visit. intensive mobbing at the nest. Twenty minutes after the The parentsreturned separatelyonce more before I ceased crow left the area, both parentsreturned, uttering soft near watching at 12:OO. calls (Balda and Bateman 1971), to which the fledglings The parents’ behavior changed in two ways after they vocally responded with various squawks, whistles, and discovered the predator. (1) Each remaining young was begging.The young were, in this manner, lead away from fed twice as long after the predator located the nest (before the nest. predation: K = 9.32 s/h, n = 3 observation periods; after The precedingobservations bring out an adaptive aspect predation: K = 18.78 s/h, n = 3 observation periods; P = of mutual parent-young communication that was dem- 0.13, Mann-Whitney U test). This was due to the male onstrated in this population by McArthur (1982). The feeding slightly more per visit after predation (median = harsh squawk given by the young jays apparently func- 88.0 s. n = 7 visits) than before uredation (median = 7 1.O tioned as a distresscall that alerted parents, other Pinyon s, n =6‘ visits; P = 0.33, Mann-WhitneyU‘ test), and to Jays, and even Steller’s Jays to the presenceof a predator parents bringing food to the nest more often following at the nest. This resulted in mobbing which, at least tem- predation (median = 1.33 visits/h, n = 3 observation pe- porarily, delayed predation. The harsh beggingquay of the riods) than before predation (median = 1.25 visits/h, n = parents may have signaled the young to leave the nest, 3 observation periods; P = 0.04, Mann-Whitney U test). although it may also be a displacement behavior early in (2) The parents remained in the nest area for long periods the nestingcycle (Balda and Bateman 1973). Normal mob- between visits to the nest. I define “nest area” as the area bing by parents does not promote suchbehavior; instead, where I could vocally or visually locate the pair from the it causesthe young to crouch and remain motionless and blind (a circle of approximately 200-m radius around the silent in the nest (Balda and Balda 1978). nest). Parents remained in the nest area for 11.5% of the Ensuingparental behaviors (mobbing and increasednest observationperiod before predation, but for 84.0% of the area attentiveness)increased parental fitness. The parents period following predation (P = 0.04, Mann-Whitney U were able to quickly respond to the distresscalls of their test). young, force the predator from the nest, and provide a Feeding appears to have increased in responseto pre- means of escapefor two of their remaining three young. dation, not as a result of caring for fewer, or older young. It is significantthat one of theseyoung survived to become In the study population, time spent feeding each nestling a breeder because,of 35 other fledglingsproduced in the varied inversely with brood size (r = -0.41, II = 29 nests, study flock in 1983, only three became breeders. P = 0.014). The observedfeedina rate followina oredation My resultssuggest that memory is adaptive in predation (K = 177.62 s/offspring/h), however, was above the upper events such as the one I report. Remembering the nest limit of the 99% confidence interval around the mean location, as well as its status(empty or not), would aid the feeding rate at other nests with three similar-aged young predator in its return to a previously located, but not emp- (K = 19.8 + 5.9 s/offspring/h, n = 7 nests). At nests that tied, nest. Remembering that a predator, who is likely to naturally had three young, I never saw males feed their return, located the nest would be adaptive for parents offspring as long per visit as at this nest (K = 14.52 becausechanging their behavior may increasetheir fledg- s/offspring,SD = 4.84 s, n = 6 nests),and I observedonly ing success. one female to exceed the feeding duration of the female This researchwas supportedin part by a grant from the at this nest (K = 6.83 s/offspring,SD = 3.3 1 s, n = 5 nests). Frank M. Chapman Memorial Fund. R. P. Balda, D. Cac- More attentivenessalso was inconsistentwith reduced camise, and S. Rothstein suggestedimprovements in ear- brood size or ageingof the nestlings.Normally, attentive- lier drafts of the manuscript. ness of smaller broods is greater than large broods (r = -0.36, n = 32 nests, P = 0.02); however, I never before LITERATURE CITED observedparents remaining for long periods in the nesting area after leaving the nest. The observed percent atten- BALDA,R. P., AND J. BALDA. 1978. The care of young tiveness(84%) was not within the 99% confidenceinterval Pinyon Jays and their integration into the flock. J. for attentivenessat nests with three young (0.0%25.6%, Omithol. 119:146-171. n = 8 nests), or for attentivenessof the entire population BALDA,R. P., AND G. C. BATEMAN. 1971. Flocking and (6.4%24.8%, n = 32 nests). Attentiveness typically de- annual cycle of the Pinyon Jay, Gymnorhinuscyano- creasesas nestlings age from 15 daysto fledging(r = -0.46, cephalus.Condor 731287-302. n = 55 days, P = 0.000). BALDA,R. P., ANDG. C. BATEMAN. 1973. Unusual mob- The jays’ increasedattentiveness helped them to detect bing behavior by incubatingPinyon Jays.Condor 75: a crow (presumably the same individual) when it came 25 l-252. the next day. The crow entered the nest tree at 11:06, and BERGER,L. R., AND J. D. LIGON. 1977. Vocal commu- after 2 min it grabbed at one nestling, which gave one nication and individual recognitionin the Pinyon Jay. harsh squawk. Immediately thereafter, the parents gave Anim. Behav. 251567-584. rack and racka calls (Bergerand Ligon 1977), in response BROWN,J. L. 1964. The integration of agonisticbehavior to which the crow left the nest without removing any in the Steller’s Jay Cyanocittastelleri (Gmelin). Univ. young. The parents followed the crow and continued to Calif. Publ. Zool. 60:223-328. mob it for 5 min. The crow returned to the nest tree 12 CLARK,L., AND D. GABALDON.1979. Nest desertion by min later, attempting to seize another nestling. Again I the Pinyon Jay. Auk 96~796-798. heard racka calls. Upon hearing the vocalizations of the MARZLUFF,J. M. 1983. Factorsinfluencing breeding suc- young, the parents and two other Pinyon Jays, flew into cessand behavior at the nest in Pinyon Jays.Unpubl. the nest tree and violently mobbed the crow, occasionally M.Sc. thesis, Northern Arizona Univ., Flagstaff, AZ. SHORT COMMUNICATIONS 561

MCARTHLJR,P. D. 1982. Mechanisms and development Departmentof BiologicalSciences, North Arizona Univer- of parent-youngvocal recognition in the Pinyon Jay sity, Flagstax Arizona 86011. Received 2 1 January 1985. (Gymnorhinuscyanocephalus). Anim. Behav. 30:62- Final acceptance15 July 1985. 14.

The Condor87:561-562 being fed by other passerine species,but has seen none 0 The CooperOrnithological Society 1985 associatedwith SageThrashers. We experimentally parasitized SageThrashers to deter- SAGE THRASHERS REJECT mine their responsesto cowbird eggs.We wanted to de- termine whether few casesof parasitism are reported be- COWBIRD EGGS causeSage Thrashers are not parasitizedor, in part at least, becausethey remove cowbird eggsbefore observersfind them. The differing appearances of cowbird and Sage TERRELL RICH Thrasher eggsindicate that a bird could distinguish be- AND tween them easily. Cowbird eggsare white with numerous small brown and gray spots, whereas Sage Thrasher eggs STEPHEN I. ROTHSTEIN are blue-greenwith red-brown blotches. We located SageThrasher nests in basin big sagebrush (Artemisia tridentatatridentata) habitat in Blaine County, Actual and potential hosts of brood parasitism by Brown- Idaho, during April and May, 1984. Nest contents were headedCowbirds (Molothrus ater) can be classifiedas either maninulated between 08:OOand 14:O0.with most manio- rejecters or accepters of cowbird eggs (Rothstein 1975, ulations being performed between 09:OOand 10:OO.At 1982a,b). Few North American speciesshow intermediate each nest, we exchangedor added an artificial cowbird egg responses(Rothstein 1982a). Rothstein (1982b) reasoned quickly and then left the area soas to minimize disturbance that once the rejection behavior appears in a ‘species, it of adult thrashers.These eggswere made of plasterof Paris has such a high adaptive value that it is rapidly fixed. and measured 21.1 x 16.3 mm. They were identical to Whatever the cause,this dichotomy betweenaccepters and eggsin Rothstein’s (1975) studyand closelyresembled real rejectersmakes it possibleto determine the responseof a cowbird eggsfound in southernIdaho (Rich, pers.observ.). given speciesby experimentally manipulatingthe contents Nests were checkedbetween 1 h and severaldays after the of relatively few nests. manipulation to determine responses.Unless noted oth- It is desirableto continue accumulatingevidence on the erwise, all nests were subjectedto only one experimental responsesto brood parasitism of untestedspecies because manipulation. no absolute criteria that explain why some species are In ten nests, we removed a thrasher eggand replacedit rejecters and others accepters have yet been identified with an artificial cowbird egg. Five of these nests were (Rothstein 1975). Although Rothstein (1975) failed to find found during egg-laying,and five were found during in- a strong relationship between taxonomy and responseto cubation. Although most natural cowbird parasitism oc- non-mimetic parasitic eggs, knowledge of the host re- cursduring the host’s egg-layingperiod, nest stagehas little sponsein all members of a family would be usefulbecause or no relation to responsein most rejecter species(Roth- many variablesrelated to morphology,behavior, and evo- stein 1976, 1977), a trend also indicated by our results. lutionary history would be somewhat controlled. Nine of the 10 eggswere ejected. The tenth egg remained The Mimidae have severalfeatures that make the family in the nest, which was deserted.The last nest was the only a good group for study of responseto brood parasitism. one where there had been only a singleegg laid at the time In particular, they have at least four of the six character- of manipulation. At an eleventh nest, we added an artificial istics suggestedby Rothstein (1975) as contributing to the cowbird egg to a clutch of four eggs. This egg was also formation of rejection behavior. Their eggsare unlike cow- ejected along with two thrasher eggs, and the nest was bird eggs,their beak is large, their nest is large and easily subsequentlydeserted. found, and they practicegood nest sanitation (Bent 1948). Cowbird eggswere usually ejected quickly, but not al- Yet, three North American mimids accept eggs of the waysimmediately; artificial cowbird eggswere still present Brown-headed Cowbird: Northern Mockingbird (Mimus at one nest after 2 h and at another after 3 h. The earliest polyglottos;Rothstein 1975) Le Conte’s Thrasher (Tox- known ejections occurred within 1, 2, and 3 h, and two ostomalecontei), and California Thrasher (T. redivivum; within 4 h, althoughin no casedid we watch a bird remove Rothstein, pers. observ.). Among North American mim- an egg. Only one of 17 ejected eggswas found, and that ids, three specieseject cowbird eggsfrom their nests:Gray was at a distance of 3 m from the nest. The egg bore no Catbird (Dumetella carolinensis).Brown Thrasher (T. ru- evidence of pecking, thereby matching previous findings fim; Rothstein 1975, 1982a), and Crissal Thrasher (T. that most speciesthat eject eggsdo so by holding eggsin dorsale;Finch 1982). A neotropical mimid, the Chalk- their bills, rather than by spiking them (Rothstein 1975). browed Mockingbird (Mimus saturninus),is also known We performed nine additional experiments with two to eject parasitic eggs(Mason 1980, Fraga 1982). types of eggs intermediate between cowbird and Sage The Sage Thrasher (Oreoscoptesmontanus) may have Thrasher eggsto get some indication of the factors that accepteda cowbird egg in the only known case of para- SageThrashers use to distinguishamong eggtypes. Single, sitism (Friedmann 1963), but most of the evidence sug- real, thrasher eggsthat were painted to resemble cowbird geststhat this speciesmay be a rejecter. Rich (1978) found eggswere ejected from four nests. These results suggest no cowbird eggs in 21 Sage Thrasher nests in an area that the difference in size between thrasher (24.8 x 16.8 frequentedby cowbirds,where at leasttwo other sympatric mm, Bent 1948, p. 429) and cowbird eggs(21.8 x 16.8 specieswere parasitized. Also, Rich (pers. observ.) has mm, Bent 1958, p. 451) is not a necessaryreleaser for examined about 40 other thrasher nests in Bingham and rejection behavior. Blame counties, Idaho, and found neither cowbird eggs Artificial eggsidentical in size to the artificial cowbird nor nestlings.In Mono County, California, both cowbirds eggs,but colored an immaculate blue (identical to eggtype and Sage Thrashers are locally common and sometimes “s” in Rothstein 1982a).were accemedat two offive nests. forage at the same horse corrals (Rothstein et al. 1980). Birds at three of thesenests were incubatingand had eject- Rothstein has seen a large number of fledgling cowbirds ed artificial cowbird eggs44 to 48 h before the blue egg