Downy Woodpecker Predation at Goldenrod Galls
Total Page:16
File Type:pdf, Size:1020Kb
j. Field Ornithol., 56(1):56-64 DOWNY WOODPECKER PREDATION AT GOLDENROD GALLS BY JOHN L. CONFERAND PETER PAICOS In Tompkins County, New York, the Canadiangoldenrod (Solidago canadensis)is common. Several gall-inducing insects lay eggseither ex- clusivelyor almostexclusively in this species.The larvae of two moth speciesinduce elongated galls after eggs are laid into the tip of the growingplant (Uhler 1951). The larvaeof a smallfly (Eurostasolidaginis) induceformation of a round gall, a "ball gall." While the DownyWood- pecker(Picoides pubescens) eats larvae of both mothsand the gall fly, the gall fly is more abundantand is the centralplayer in the followingstory. The gall fly larvagrows during the summerinto a maggotthat weighs about .04 g by fall. Larvae pupatein late winter and adultsemerge in early spring,repeating the cycleannually. Fly larvae may be parasitized by either of two wasplarvae (Eurytomaobtusiventris or E. gigantea)or preyedupon by a beetlelarva (Mordellistenaunicolor). On averagethese insectlarvae weigh about .005 g. Downy Woodpeckersuse all these larvae and pupaeas a food resource.To get a larva or pupa a Downy Woodpeckerwill chisela distinctive,conical hole into the side of the gall. In one surveyas manyas 40% of the gall insectswere eatenby the woodpecker(Milne 1940). Yet, in another surveyconducted in Tomp- kins County, New York (Uhler 1961), as few as zero to 9% were eaten from year to year. Working in the samearea as Uhler, we found areas with nearly 50% of the gall insectseaten, yet other areaswith virtually no insectseaten. This preliminaryinformation suggested that gallinsects could be an important food for Downy Woodpeckersand that some factorscould greatly alter the accessibilityof the gall insectsfor the woodpecker.Herein, we report on the influenceof severalfactors on predationby the DownyWoodpecker on the goldenrodball gall. These include:gall height and diameter,patchiness of the galls,distance from the gall to woodycover, snowdepth, and the woodpecker'sselection of gallsdependent on the speciesof insectin the ball gall. METHODS We sampledabandoned farmland in Tompkins County, New York from Fall 1980 to Spring 1983. Goldenrod occurred in patcheswith densitiesfrom lessthan 1 to about100 stems/m2. Average stem densities approximated10 to 20 stems/m2 with aboutone out of 5 stemshaving one or more galls.The goldenrodgall fly and its insectparasites and predatorswere identifiedby meansof the helpful work of Uhler (1951, 1961). Surveysof seasonalpredation rates were conductedby repeated searchesfor taggedstems with galls.Strips of labelingtape were num- bered and stapledaround the plant stembelow the gall. The tagsre- sembled leaves on the stem and were not noted to have an influence on 56 Vol.•i, •o.I DownyWoodpeckers andGoldenrod [57 Downy Woodpeckerpredation. Tags aided relocationof the stemson later visitsand reduced subjectivity in selectionof gallsfor the repeated surveys.Some tags were lost,and in mid-wintermany stems with galls were buried by snow.In one surveythe percentrediscovery 4 months after taggingand after snowmelt was98% (513 out of 524). Normally 100% of the gallsfound on one trip couldbe rediscoveredon the next trip if there wasno interveningsnowfall. Galls less than .5 mm diameter were ignoredsince these galls rarely had larvaein them and rarely were attackedby Downy Woodpeckers. The effect of distancefrom the gall to woody cover (stems•3 m height)on the frequencyof predationwas measured. For all but one of the studysites the transitionfrom woodycover to field wasabrupt and formed a straightline. In the one exception,the 1980-1981 survey, patchesof woodyvegetation occurred irregularly throughoutthe site. At thissite, gall diameter, height, distance to woodycover and patchiness were measuredconcurrently. Patchiness was estimated by the average distanceto the 5 nearestgalls. We suggestthat Downy Woodpeckersselect directly f•r galls con- tainingthe larger fly in preferenceto gallscontaining the smallerwasps or beetle.This selectionwould involvethe followinginsect life histories (from Uhler 1951, 1961) and Downy Woodpeckerbehavior. During October the fly larva, usingits mouth hooks,prepares an exit tunnel up to but not includingthe epidermis.In May the adults,which lack mouth hooks,push open the epidermallayer and emerge.Downy Woodpeckersfrequently enlarge this exit tunnel when extractingthe larva(Moeller and Thogerson 1978). However, if the fly larvais preyed upon or parasitizedby a larva of other insects,this happensin the summerbefore the fly larva preparesan exit tunnel. The other insect larvaedo not makea tunnelbecause adults of the other speciesget out on their own effort. Sometimesa Downy Woodpeckerwill make a few scatteredpeck markson a gall and then abandonit. We hypothesized that the woodpeckerwould be more likely to abandona gall, leaving peck marksbut no chiseledhole, if it couldnot find an exit tunnel. To testthis hypothesis nearly 3000 gallswere collected in March from 3 locationsin 2 yearswithout any examinationfor DownyWoodpecker peckmarks or chiseledholes. Since the 3 samplesyielded the sametrend, they were pooled. These sampleswere examinedmeticulously in the laboratoryunder good light conditions.A total of 1077 gallswere found with the hard-to-seepeck marks or large, chiseledholes. Although some peck marksmay have been missed,this possibleerror would occurwith equalprobability for gallscontaining a fly or containingthe other insects and would not biasthe results.The attackedgalls were openedto de- termine the insectoccupant. The proportionof gallswith peck marks but no chiseledhole wascalculated for gallswith a fly larva or with a carnivorous larva. The total predation by Downy Woodpeckersover a large area was estimated.A field next to a suetstation was surveyed to determinethe 58] j. L. Conferand P. Paicos j. FieldOrnithol. Winter 1985 number of gallsopened by woodpeckers.Sixty-four 1 m2 sampleplots were positionedalong a grid that covereda 1 ha field. The surveywas conductedin April 1981 when the seasonalpredation was finished. Similar fields within about 300 m of the suet station compriseda total area with goldenrodgalls of about5 ha. Woodyshrubs were distributed throughout the surveyedand adjacentareas and Downy Woodpecker predation occurredthroughout. The assessmentof predation by Downy Woodpeckersis complicated by Black-cappedChickadees (Parus atricapillus) which also peck holes into these galls. Downy Woodpeckerstend to make a tidy, narrow, conicalhole by pecking,while Black-cappedChickadees tend to make a messy,large, irregular hole by grabbingbits of the gall with their bill and tuggingthem free. Schlicter(1978) found up to 50% of the galls attackedby the chickadeein southernOntario. In the Ithacaarea Black- cappedChickadee predation is rare. Uhler (1961) never noticed this out of 17,000 galls.J. Glaseand D. Gray, who collectivelyspent about 200 weeksin field studiesof chickadees,saw this predation about 5 times (pers.comm.), while we observeda chickadeeattacking a gall once.Out of thousandsof gallswe have examined,probably fewer than 30 were attackedby the Black-cappedChickadee and few of theseoccurred in our quantifiedsurveys. In our analyseswe have made the slightlyer- roneousassumption that all holeswere made by the DownyWoodpecker. RESULTS The DownyWoodpecker fed on gallsfrom early Novemberuntil late April in a highlysporadic manner (Table 1). Gallstagged for the 1980- 1981 surveywere buried by snow from January to mid-February.A thaw then exposedthe gallsand wasfollowed by a burst of predation detectedon the mid-Februarysurvey. Two other surveysduring 1980- 1981 from widelyseparated areas were lessintensive, but still indicated a pulseof predation followingthis thaw. During subsequentweeks of 1980-1981 the predationrate variedgreatly and wasnot relatedto any environmental condition that we noted. The 1981-1982a surveyinvolved a randomset of gallswith about% of the galls 15-30 m out from a forest edge. Characteristicsof these galls were similar to the precedingset of galls, and snow cover was similar for these 2 years. For both of the above surveys,almost all of the predation occurredin early spring. Gallsin the 1981-1982b surveywere unique in that theywere selected to be on tall stems(> 1 m) and closeto trees (<2 m). Unlike gallsof other surveysnone of thesegalls wasburied by snow.Almost all the predationon this latter set of gallsoccurred in fall, with virtually no gallseaten in spring.A DownyWoodpecker was twice seen in this area during late spring,so the absenceof predationwas not likely due to the absenceof a bird. Casualobservations of other patchesof gallssuggest that heavypredation on large gallson tall stemsnear treesbefore Jan- uary is a commonevent. Vol.•6, No. I DownyWoodpeckers andGoldenrod [59 60] j. L. Conferand P. Paicos j. FieldOrnithol. Winter 1985 TABLE2. Predationby the DownyWoodpecker on goldenrodball gallsas influenced by the distancefrom woodyvegetation. Survey Distance" Number date (m) Eaten (%) Not eaten Spring 1982 0-5 23 (14) 128 15-20 6 (4) 130 9.21 25-30 4 (3) 140 Fall 1982 0-3 19 (16) 103 3-10 4 (5) 85 22.27 6-10 0 (0) 109 Spring 1983 o-3 27 (24) 85 3-6 8 (5) 158 24.03 6-10 8 (10) 73 Spring 1984 0-2 8 (9) 78 10-12 5 (2.5) 194 11.98 2o-22 3 (3) 105 Distanceintervals varied from surveyto surveydepending on habitatavailability. Three surveyswere conducted during 1982-1983 in widelyseparated areasand involveddifferent birds.For all 3 of thesesurveys most of the predationoccurred early in winter, whichdiffers from the 1980-1981