SEASONAL CHANGES IN FEEDING SUCCESS, ACTIVITY PATTERNS, AND WEIGHTS OF NONBREEDING SANDERLINGS (CALIDRIS ALBA) J. L. MARON1'2 AND J. P. MYERS1,3 •BodegaMarine Laboratory,P.O. Box247, BodegaBay, California 94923 USA; 2Departmentof Biology,University of North Dakota,Grand Forks, North Dakota58202 USA; and 'The Academyof Natural Sciences,19th and the Parkway,Philadelphia, Pennsylvania 19103 USA A•3STl•ACT.--Seasonalchanges in Sanderling(Calidris alba) feeding success,time budgets, and weights were followed throughoutthe nonbreedingseason at BodegaBay, California. Sanderlingsspent more time roosting in fall than in winter, and in fall adults spent more time roostingthan juveniles.Sanderling prey capturerates were high in fall and springand declinedthrough winter. Sanderlingweights paralleled seasonal changes in feedingsuccess and activity patterns:birds were heaviestin fall and spring and lightest in winter. These resultsreflect lowered food availability in winter and imply that birds may have difficulty balancingtheir energy budgetsduring part of the nonbreeding season.Received 6 July 1984, accepted31 December1984. RESOURCEconditions faced by northern Tem- beachesand on tidal sandflats. During winter, perate Zone shorebirds often deteriorate con- stormsbeset the beachesand drasticallyreduce siderablythrough the nonbreedingseason. Prey prey numbers. Cumulative predation on sand availability may decline due to reductions in flats also reducesprey availability (Myers et al. prey density (Evans 1976, Evans et al. 1979, MS). In spring the recruitmentof new prey co- Goss-Custard1980, Myers et al. 1981, Quam- horts,particularly Emerita analoga, increases prey men 1980), changesin physical conditions of availability on beachesonce again (Connorset the foraging environment (Dobinsonand Rich- al. 1981). In this paper we consider whether ards 1964, Myers et al. 1981, Pienkowski 1982), these fluctuationsin prey abundanceaffect lo- or alterations in prey behavior (Smith 1975, cal Sanderling prey capture rates,activity pat- Reading and McGrorty 1978, Evans 1979, Pien- terns, and weights. kowski 1983). The question is whether these alterations in STUDY AREA AND METHODS prey availability have material effectson shore- bird behavior. This need not be the case if the The studywas conducted from August1982 to April fluctuationsin availability occur over a range 1983 at BodegaBay, which is along the central coast of values where caloric intake rates are not af- of California (Connors et al. 1981). Reference pho- fected-for example, in the asymptoticpart of tographsare archivedin VIREO (Academyof Natural Sciences). the curve relating foraging rate to prey density. Sanderlingsarrive at BodegaBay in July and Au- Fluctuationswithin this range will have neg- gustand departin April and May (Myers 1980,Myers ligible impact on shorebirds. et al. 1985).Birds at BodegaBay feed at two exposed If fact, many shorebirdsdo alter their behav- sandy beachesduring high tide and then fly to a ior on a seasonalbasis in relation to prey avail- nearby tidal sand flat to feed as the tide recedes(Con- ability. They spendmore time feedingas winter nors et al. 1981). progresses(Goss-Custard 1969, 1977; Goss-Cus- On beaches,Sanderlings feed mostly on crusta- tard et al. 1977; Puttick 1979) and switch to new ceans,primarily the sandcrabEmerita analoga and the feeding sites or new prey (Goss-Custard1969, isopod Excirolanalinguifrons (Connors et al. unpubl. Smith 1975, Evans 1976, Pienkowski 1982). data; this study). Less frequently, Sanderlings feed on small polychaetesor insectsand talitrid amphi- Sanderlings(Calidris alba) wintering at Bo- pods, often from areas in and around beach wrack dega Bay in central coastalCalifornia experi- (Yaninek 1980, pers. obs.).On tidal sand flats, Sand- ence a regular annual cycle in prey abundance erlings feed on molluscs,small crustaceans,and var- (Myers et al. 1981, unpubl. data; Connors et al. ious polychaetes(Couch 1966, Recher 1966, Myers 1981). Upon arrival and until midautumn, re- unpubl. data; this study). sources are abundant both on outer sandy Individualactivity patterns and feeding rates.--Season- 580 The Auk 102: 580-586. July 1985 July1985] WinteringSanderlings 581 al changesin Sanderlingfeeding behavior and prey the sameapproach when analyzing capture rates on capturerates were assessedby focalanimal sampling tidal flats, for which, however, we comparedonly (Altmann 1974).We followedfocal birds at high tide distributions within the same range of tidal eleva- (1.0-2.0 m) on beachesand at low tide (-0.3 to +0.9 tions. Prey density within our study area varies as a m) on tidal flats. For behavioral observations, we function of tidal elevation both on beaches and tidal walked randomlyselected stretches of beachor tidal flats, but Sanderlings use the beachesonly over a flat and, in a given group of birds, observedthe restricted range of tidal elevations (Connors et al. bandedindividual next to the firstbanded bird sight- 1981). ed. We followedeach focal animal for as long as it Populationactivity patterns.--We measured the per- wasin sight, but not longer than 10 min, recording centageof time Sanderlingsspent roosting vs. feed- behavior(with a taperecorder) every 30 s. The mean ing using instantaneousscan sampling (Altmann observation time per individual was 7.8 min. After 1974). To do this we recorded the behavior of birds eachobservation period, another focal bird was cho- as soonas they were spottedas we walked along the sen. In this way, different individualsfeeding in a entire length of the two beachesand tidal flat. In variety of social conditions were observed. We re- most casesactivity patterns were noted during the cordedprey capture continuously throughout the ob- courseof regular censuses.Scan sampleswere con- servationperiod and identified prey itemswhenever ductedat all timesof day and acrossa wide rangeof possible. tide heightsand tidal elevationsin all monthsduring Prey identification was simple on outer beaches the study. (Tide height is used in its conventional due to the markedvertical zonation of differentprey sense.Tidal elevation, constantthrough time for a (Myers 1979,Connors et al. 1981,Myers et al. 1981, given site, is the absolutevertical distancebetween pets.obs.). Prey identificationon sand flats was more a given positionon the sandflat and mean lower low difficult.Prey taken could be classified as either poly- water, the standardreference for tide height.) Most chaetes or "unknown." Observations were made us- scansamples were taken only once per day in any ing a 15-60 x spottingscope. We did not selectroost- given habitat;when they were taken more frequent- ing birds as focal animals and discontinued ly they alwayswere separatedby at least 1 h. observationsof focalindividuals if they beganroost- On 9 daysin Septemberand October(when juve- ing. niles and adults could be distinguishedon the basis Many individualswere sampledseveral times dur- of plumage differences),we determined the percent- ing the season.On the beach,we sampledthe same ageof juvenilesand adultsthat were roostingat high individual in more than 1 month 28% of the time tide. and in more than 3 months 5% of the time. On tidal To determinewhether birds spentmore daylight flats, we sampledthe sameindividual in more than hours feeding in winter vs. fall, we calculatedthe 1 month 21% of the time and in more than 3 months mean number of daylight hours in fall and winter 3% of the time. This overlapamong sample periods, and multipliedthese values by the meanpercentage a result of the randomsampling procedure used to of daylighttime birdsspent feeding within thesesea- select focal animals, reduces the likelihood that dif- sons. The mean percent daylight time birds spent ferencesin foragingrates among sample periods were feeding within a seasonwas determined by calculat- due to differencesamong individuals. ing the percent of birds on beachesand tidal flats We estimatedtidal elevation of the focalbird every that werefeeding at a giventime and averagingthese 30 s by recordingthe bird's positionrelative to per- valuesfrom observationsat all timesof day (for anal- manent stakes of known tidal elevation. When the ysis of activity patternson beaches)and all tidal el- focal bird was not near a stake of known tidal ele- evations(for analysisof activitypatterns on tidal flats). vationbut wasfeeding at the tide line, tidal elevation To determine how time of day and tidal .height at thatlocation was calculated arithmetically (for de- influencedthe amount of time birds spent roosting, tails seeMaron 1984).Data on birds feedingat or we performeda multiple regressionusing hour of near the tide line were gatheredwithin 45 min of day and tidal height asindependent variables to pre- the estimated low tide, to control for short-term vari- dict percentage of time spent roosting. For the ation in feeding successdue to changesin prey be- regressionwe transformedthe time-of-day variable havioror in substratepenetrability caused by reced- into hoursbefore or after 1400.The regressionanal- ing tide and drying substrate(Vader 1964, Smith 1975, ysis was performed separatelyfor beach and tidal- Pienkowski1983). We systematicallyand regularly flat data. altered our observationposition to avoid site or tidal- Weights.--Wecaught Sanderlingseach month by elevation biases. placingmist nets around roostson beachesat night. To testfor seasonaldifferences in prey capturerates After capture, birds were weighed, color-banded, on exposedbeaches, we comparedfrequency distri- measured(tarsus