j. RaptorRes'. 37(2):117-124 ¸ 2003 The Raptor ResearchFoundation, Inc.

BREEDING SEASON HABITAT USE AND ECOLOGY OF MALE NORTHERN PYGMY-

ALAN R. GIESE 1 Departmentof Fisheries and Wildlife, State University, Corvallis, OR 97551 U.S.A.

Emc D. FORSMAN U.S.Forest Service, Research Station, 3200 SWJeffersonWay, Corvallis, OR 97331 U.S.A.

ABSTRACT.--Weexamined habitat use,nest location,diet, and activitypatterns of radio-markedNorthern Pygmy-Owls(Glaucidium gnoma) during four breedingseasons (1994-97) in fragmentedforests on the OlympicPeninsula, . We observedforaging in all availablevegetation categories, but patterns of use were non-random.Structurally diverse and older forestswere most heavilyused, openingsand patchesof saplingsreceived the leastuse, and use of edge wasintermediate. We locatedeight nests,all in cavitiesin patches of structurally-diversetbrest. We found no clear evidence that nests were locatednear edges.Northern Pygmy-Owlswere diurnally active,and male owlsdelivered food to femalesduring nest establishmentand incubation,and to both femalesand nestlingsduring brooding. Date of fiedging varied tkom mid-Juneto mid-July.Fledging was synchronous,and minimum estimates of brood size ranged from 1-5. Diet included a mix of small , and insects.Our results suggestthat the creationof openingsby clear-cutlogging is unlikely to benefit Northern Pygmy-Owls, and that the replacementof structurallydiverse forests with uniform forestsmay be detrimental. KEYWOP, DS: NorthernPygmy-; Glaucidium gnoma;fragmentation; habitat use.

ESTACION REPRODUCTIVA, USO DE HABITAT Y ECOLOGIA DEL BUHO PIGMEO MACHO DEL NORTE

RESUMEN.•Examinamosel uso de hibitat, la localizaci6nde nidos,la dietay los patronesde la actividadde bfihos pigmeos(Glaucidium gnoma), durante cuatro estacionesreproductoras (1994-97) en bosquesfrag- mentadosen la peninsulaOlympic, en Washington.Observamos el forrajeo en todaslas categoriasde ve- getaci6ndisponibles, cuyos patrones de uso no estabandeterminados pot el azar.Los bosquesmils viejos y estructuralmentediversos fileron los m•s usados,los clarosy parchesde •rboles fueron los mucho menos usados,la utilizaci6nde bordesfue intermedio.Localizamos ocho nidos,todos en cavidadesde carpinteros en parchesde bosquesestructuralmente diversos. No encontramosninguna evidenciaclara que los nidos estuvieransituados cerca de los bordes.Los bfihospigmeos del norte fueron mis activosdurante el dia, los bfihosmachos entregaron el alimentoalas hembrasdurante el establecimientode nidosy la incubaci6n, tanto a hembrasy pichonesdurante el empollamiento.La fecha de crecimientode plumas en que los polluelosdeNan abandonarel nido vari6 a parfir de mediadosde junio hastamediados de julio. E1 creci- miento de plumasfue sincr6nicoy las estimacionesminimas del tamafiode la nidadafiuctuaron entre 1 y 5. La dieta incluy6una mezclade pijaros,mamifkros e insectospequefios. Nuestros resultados sugieren que espoco probableque la creaci6nde clarospot la tala beneficieal bfiho pigmeodel norte y que el reemplazo de bosquesestructuralmente diversos pot bosquesuniformes puede set perjudicial. [Traducci6n de C6sar Mirquez]

Holt et al. (1990) labeled the Northern Pygmy- westernforests remains scant,and published infor- Owl (Glaucidium gnoma) 's least- mation is often equivocal.Yet western forestshave studied owl. Thirteen years later, detailed infor- experienced decades of intense resource extrac- mation for this locally-common inhabitant of tion pressure(Parry et al. 1983) and it is presently impossibleto estimatehow suchchanges might af- • Presentaddress: Department of Biology,Arizona State fect Northern Pygmy-Owls.Here, we report on University,Tempe, AZ 85287-4601U.S.A.; e-mail address: Northern Pygmy-Owl habitat use and ecology in [email protected] forestsheavily fragmented by clear-cutlogging.

117 118 GmsEAND FORSMAN VOL. 37, NO. 2

The range of the Northern Pygmy-Owlextends many owls each day as possible(typically two to eight) from Alaskato Central America, includingforests and beginningwhere we had finished the day before Thus, eachowl was located at all timesof the day,because from the west coast to the a new owl typicallytopped the list each day. Location (Johnsgard1988). They are diurnally active,use times ranged from about an hour before sunriseto about cavitynests, and are generalistpredators of small an hour after sunset. birds,mammals, and insects(Bent 1938,Johnsgard We classifiedhabitat into five vegetationcategories 1988). Northern Pygmy-Owlshave been commonly basedon structuralcharacteristics. The Open-Saplingcat- egoryincluded areas dominated by smallconifer saplings sightedin or near openings,leading to statements (1-4 m tall) intermixed with extensiveareas (>25% cov- that they preferentiallyforaged in openings(Bent er) of bare ground and shrub cover,mostly on recent 1938, AOU 1983, Johnsgard 1988), nested near clear-cuts.The EarlyStem Exclusion category was dominat- edges (Webb 1982, Reynolds et al. 1989), and ed by youngconifers (5-15 m tall) with few openingsin the overstoryand with dense, overlappinglimbs in the might benefit from partial forestclearing (Johns- understory.The Late StemExclusion category was domi- gard 1988). In contrast,others have suggested that natedby medium-sizedconifers (typically 20-40 cm DBH partial forestclearing may be detrimentalto North- and >15 m tall) with a closedcanopy and open under- ern Pygmy-Owls(Marshall 1992). story.The StructurallyDiverse category was characterized Hayward and Garton (1988) used call-response by a wide varietyof sizesand a multilayeredcanopy. It consistedprimarily of matureand older forests(typi- surveysto examine resourcepartitioning by small cally >80 yr), but alsoincluded mid-aged stands charac- forest owls in and concluded that the North- terizedby high structuraldiversity. Finally, the Edgecat- ern Pygmy-Owlwas a habitat generalist.Although egory included all areas within 30 m of an intersection their workis the mostextensive study to date,many between a patch of Open-Sapling,and any of the other questionsremain about habitat useby this . categories.Thirty meterswas used because pilot-study ob- servationssuggested that foraging flights by Northern We usedradiotelemetry to studyhabitat use for two Pygmy-Owlsrarely exceededthis distance.A foraging behaviors,foraging and nesting.We were specifi- flight wasdefined as any flight which includedor ended callyinterested in whetherpartial forest clearing with an attempt to capture prey. Becausepatchiness in m•ght benefit this species.We alsodescribe home the studyarea waspredominantly the result of clear-cut range size, diet, activitypatterns, nest characteris- logging, patcheswere generallyeasy to delineateand classify.We delineatedpatch boundariesusing ARC/ tics,and nestingbehavior. For the purposesof this INFO (ESRI, Redlands,CA) and digital orthophoto- paper,we hereafteruse 'Northern Pygmy-Owl'and graphs.We visitedeach patch and basedvegetation clas- 'owl' synonymously. sifications for both used and available cover on visual inspection. METHODS We usedcompositional analysis (Aebischer et al. 1993) The studyarea was a mixture of federal,state, and pri- and programRESELECT (Leban 1994; available at http. vate landson the northwestcorner of the OlympicPen- //ces.iisc.ernet.in/hpg/envis/resdocl 120.html) to com- •nsula,Washington. The area was hilly to mountainous pare relativeuse among vegetation categories. Composi- with elevationsranging from 50-1350 m. Mean annual tional analysistreats the individualas the samplingunit, precipitationranged from 150-250 cm. Natural vegeta- accountsfor the unit-sumconstraint of proportions,and tion wasdominated by coniferousforests of westernhem- allowsunique availabilityfor each individual.We usedan lock (Tsugaheterophylla), Douglas-fir ( menzie- F-testto determine study-widedeviation from random us- szz),western red-cedar ( plicata),and silverfir (Abies age,and pairwiset-tests for differentialuse between veg- amabilis). Valley bottoms typically included variable etation categories. amountsof bigleaf (Acermacrophyllum) and red al- Use ratios were calculatedby dividing proportionate der (Alnusrubra) (Hendersonet al. 1989). Even-aged use by proportionate availability.Proportionate use was patchesof primarily Douglas-firresulted from logging definedas the proportionof locationsfor a givenowl in and silviculture. a given vegetation category.All locations were deter- We locatedowls by walkinglogging roads while vocally mined by homing to an owl with a hand-held receiver •m•tatingtheir calls.Surveys were conducted during April until the owl was locatedvisually, or until triangulation and May 1994-97,and were not designedto samplethe indicated that the owl was directly overhead.Locations study area equitably. Rather, we concentrated on areas were mappedin the field on aerial photographsand sub- wherewe had observedowls previously, and surveyed oth- sequentlydigitized using digital orthophotographs.Lo- er areasless intensively. When found, owlswere captured cations of owls <50 m from an active nest were not in- •n mistnets, marked with bandsissued by U.S. Geological cluded in the analysesbecause of the possibilitythat SurveyBird Banding Laboratory;and fitted with back- thoselocations represented nest activityrather than for- pack transmitters(Model BD-2G ca. 2.5 g with harness, aging. Removal of such locationswould introduce bias if Holohil SystemsLtd., Carp, Ontario, Canada). they were foraging locations.However, because all nests Radio-marked owls were relocated several times each were locatedin the mostheavily-used vegetation category week during the 15-wktransmitter life span.We worked (see below), this bias would be conservative relative to from a sequentiallist of radio-markedowls, locating as our conclusions. We assessed the error associated with JUNE2003 PYGMY-OWLHABITAT USE 119 overhead triangulationsby mapping the location of 12 RESULTS transmittersplaced in by an independent observer. Proportionate availabilitywas defined as the propor- We radio-marked 21 owls during fbur field sea- tion of a given vegetation category within the minimum sons (1994-97), including 16 males, one fkmale, convex polygon (MCP) that encompassedthe locations and fbur sex-unknowns. We had sufficient data to for a given owl. We used the MCP for availabilitybecause estimate ranges and conduct habitat use analyses we believe it best approximated the area in which a breeding owl had the opportunity to forage.A circle cen- fbr nine males (Table 1). Of those, six nested and tered on the nest was unsatisfactorybecause our pilot fledged young, one nested and failed to fledge studyrevealed that nestswere not alwayscentrally located young, one nestedand wasthought to have failed, within territories, and territorial overlap was minimal. and the nesting statusof one was undetermined. Thus, vegetationwithin such a circle might receivelittle Radio-marked owls excluded fi'om the analysisof use due to the presence of a neighboring owl. Alterna- habitat use included fbur males with fkwer than 30 tively,kernel estimators(Seaman et al. 1998) are more biased toward heavily-usedvegetation patches than are locations (Table 1), fbur that left the study area, MGPs,and may exclude areasthat are availablebut used two that died, one whose sex was unknown, and infrequently.We evaluatedthe sufficiencyof our avail- one fkmale. Although we surveyedin consecutive ability estimateswith a post hocanalysis of MGP size in years, we never trapped any previously banded relation to the number of relocations.An inflection point owl s. wasapparent at ca. 30 locationsso we removedoMs from further analysisif they either died or left the studyarea We collected a mean of 49 locations per owl before we had accumulated 30 locations. (range = 34-66, N = 9). Estimatesof home range For comparison,we estimatedhome range sizesusing size (mean +_SE) were 296 +- 42 ha (N = 9) fbr both 100% MGP and 95% fixed kernel methods.In gen- the MCP method and 209 + 28 ha (N = 9) fbr the eral, owlsin this studyoccupied well-defined home rang- fixed kernel method (Table 1). Of all locations, es during the breeding season.However, one owl made two excursions of 6 and 11 km from its core use area for 49% were confirmed visuallyand 51% were esti- three dayseach in April, and then returned and nested mated by triangulation. Estimatedtriangulation er- near its original trapping location. We removed the six ror (mean +SE) fbr transmitters placed in trees locations collected during these excursionsfrom MGP was 11.6 -+ 2.3 m (N= 12). home range estimation,but included them in the kernel estimates.Reynolds and Linkart (1990) discussedextra- Use of vegetation categories fbr fbraging was range movementsin FlammulatedOMs (Otusfiamme01us) nonrandom (F4,s= 29.41, P < 0.01). The Structur- and Linkart et al. (1998) removed extra-range move- allyDiverse category was the most used, fbllowed by ments from their home-range determinations. Programs Late StemExclusion, Edge, Early StemExclusion, and GALHOME (Kie et al. 1996) and KERNELHR 4.28 (Sea- Open-Sapling(Table 2). Confidence in the rank as- man et al. 1998) were used for MGP and kernel analyses, respectively. signmentsfbr the Open-Saplingand Structurally Nestswere located by observingmales delivering food versecategories was high, as indicated by low P-val- to incubating females. The distance from each nest to ues fbr pairwise comparisonsof rank with other the nearest edge (dE) was measured in the field with a vegetation categories.Confidence in the relative 50-m tape and comparedto the mean distance-to-nearest- ranks of Late StemExclusion, Edge, and Early Stem edge (dr) for 100 randomlygenerated points in the same stand. Random points and associateddistances were gen- Exclusionwas low, as indicated by mostly non-sig- erated with ARC/INFO. A studentized Z-statistic was cal- nificant P-valuesfbr pairwise comparisons.Our re- culated for each nest (Eq. 1) sultssuggested a dichotomybetween 'fbrested' and 'non-fbrested' vegetation categories. We reana- z = (dE- &)/Sr (1) lyzed the data with the Open-Saplingcategory re- where Sris the standard deviation of the random point moved and fbund that the relationshipsbetween distances. We used a one-tailed t-test to test for Z < 0. the remaining categories were qualitatively un- We climbed to each nest post-fledgingand measuredcav- changed.We also analyzed 'the data with and with- ity entrance and tree dimensions.Tree heights >20 m were estimatedwith a clinometer. We collected pellets out the one owl whose nesting statuswas not con- and prey remnants from the ground near nests two to firmed and the results were again qualitatively three timesper wk, and recorded all observationsof owls unchanged. with prey. We pooled pellets and remnantsfor each col- We located eight nests,all of which were in dead lection date and nest, and estimated minimum trees in cavitiesexcavated by (Table prey counts for each pooled sample. Each pellet was treated as independent for counts of insect prey. We re- 1). Estimateddistance to the nearestedge (• + SE) corded owl behaviors during three dawn-to-dusk nest was 59 + 16 m fbr the nests and 99 - 3 m fbr the watches and 34, 2-hr focal observations. random points. The studentized diffkrence be- 120 GIESEAND FOP,SMAN Vot.. 37, No. 2

ZZZ Z JUNE 2003 PYGMY-OWLHABITAT USE 121

Table 2. Use of vegetationcategories by adult male Northern Pygmy-Owlson the OlympicPeninsula, Washington, 1996-97, expressedas logged use-ratiodifferences (SE). A positivevalue indicatesthat relativeuse for the row categoryexceeded that of column category.Categories are ranked from most (4) to least (0) used by adding the number of positiveuse-ratio differences acrossrows.

VEGETATION OPEN- EARLY STEM LATE STEM STRUCTURALLY CATEGORY SAPLING EXCLUSION EDGE EXCLUSION DIVERSE RANK

Open-Sapling - 2.02** - 2.68** - 2.74** - 3.44** 0 (0.51) (0.40) (0.47) (0.33) Early Stem 2.02** -0.66 -0.72 - 1.42' 1 Exclusion (0.51) (0.48) (0.51) (0.49) Edge 2.68** 0.66 -0.06 -0.76** 2 (0.40) (0.48) (0.35) (0.17) Late Stem 2.74** 0.72 0.06 -0.70 3 Exclusion (0.47) (0.51) (0.35) (0.31) Structurally 3.44'* 1.42* 0.76'* 0.70 4 Diverse (0.33) (0.49) (0.17) (0.31) * P < 0.05 from two-tailedt-tests for pairwisedifferences in log-ratios. ** P < 0.005 from two-tailedt-tests for pairwisedifferences in log-ratios. tween nestsand random points (• _+SE) was Z = males typically perched near the nest while males -0.51 + 0.31, and did not provide sufficient evi- foraged and delivered food. We documented date dence to showthat nestswere associatedwith edges of fledging for nine nests (Table 1). At four nests (One-tailed t7 = 1.64, P -- 0.07). All nests were we observed chicks exiting the nest, and in each located in StructurallyDiverse forest patches, and case, all known chicks from a given nest exited seven nests were in patches of late successional, within a 6 hr period. Minimum estimatesof brood (>200 yr old) coniferousforest. The eighth nest size varied from one to five based on the maximum was in a relatively young, mixed patch of - number of fledglingsobserved simultaneously (Ta- ous and deciduoustrees that had regeneratednat- ble 1). urally following logging. DISCUSSION Northern Pygmy-Owlsconsumed a variety of small birds, mammals and insects (Tables 3 and 4), We found strong evidence that patterns of use and males provisionedfemales and nestlingsdur- differed from patterns of availability,indicating ing incubation and brooding. In ca. 100 hr of nest that owlsdiscriminated between the vegetationcat- observation we no indication of females for- egorieswe defined. Although usewas concentrated aging. We observedfemales acceptingprey items in structurally-complexforests, we observedforag- from males and retrieving cached prey items, but ing flights in all vegetationcategories. Therefore, not leaving or returning with fresh prey items of locationsin seldom-usedcategories cannot be at- their own. During dawn-to-dusk observations at tributed to owlsin transit.Use of edges(as defined three nests,females were either in the nest cavity here) was proportionate to availability, the least or perched within 50 m, and radiotelemetry sug- used vegetation category consistedprimarily of re- gestedthat males visited the nest stand every 1-3 cent clear-cuts,and nestsdid not appear to be as- hr. Additionally, during the egg-layingperiod, re- sociatedwith edges.Thus, our resultssuggest that

Table 3. Percentcomposition of the diet of adult male Northern Pygmy-Owlson the OlympicPeninsula, Washing- ton, 1996-97, based on three different methods of data collection.

METHOD N MAMMALS BIRDS INSECTS TOTAL

Direct observation 59 45.8 50.8 3.4 100.0 Pellet analysis 83 59.0 18.1 22.9 100.0 Prey remnants 8 12.5 87.5 0.0 100.0 122 GiEsE •ND FORSM2tN VOL. 37, NO. 2

Table 4. Speciesof mammalsand birds identified as prey of adult male Northern Pygmy-Owlson the Olympic Peninsula,Washington, 1996-97, basedon direct observation,pellet analysisand prey remnants.

MAMMALS BraDS

Shrew (Sorexsp.) GrayJay ( Perisoreuscanadensis) Coast ( Scapanusorarius) Chestnut-backedChickadee ( Poecilerufescens) mouse (Peromyscusmaniculatus) Brown Creeper (Certhiaamericana) Red-backedvole (Clethri0n0mysgapperi) Winter Wren (Tr0gl0dytestroglodytes) (Microtussp.) Golden-crownedKinglet (Regulussatrapa) Townsend'schipmunk (Tamiastownsendii) (Catharussp.) Varied Thrush ( Ixoreus naevius) Dark-eyedJunco (Juncohyemalis) White-crowned Sparrow ( Zonotrichialeucophrys) Golden-crownedSparrow ( Zonotrichiaatricapilla)

the creation of openingsby clear-cutlogging is un- ever, our sample was small and the results were likely to benefit Northern Pygmy-Owlsand that the nearlysignificant (N = 8, P = 0.07). Furthermore, replacementof structurallydiverse forests with uni- our approach assumedthat potential nestswere form forestsmay be detrimental. This conclusion evenlydistributed in forest patches.While it would is subjectto the caveatthat habitat use may vary be useful to know the true distributionof potential temporally,and we cannot exclude the possibility nests, estimating such a distribution would be that different typesof habitat receiveheavier use problematic,and might only be relevant if poten- in different seasonsor years. tial nestswere in limited supply.Our analyseswere Owls in this studydid not use edge habitat or also sensitive to scale. The mean distance from a openingsheavily, and detectionbias is one plausi- nest to the nearest edge (59 m) may be a biologi- ble explanationfor thisinconsistency with previous callymeaningful proximity that wasnot statistically anecdotal accounts. Prior to initiating this radio- significant in our study due to an abundance of telemetry study,we observedNorthern Pygmy-Owls edges. A mean distance of 99 m from random mostfrequently in recent clear-cutsand along edg- points to the nearest edge givesan indication of es. Moreover, our visual confirmation rate was ap- the ubiquitousnessof edgesin our studyarea. proximately50% for owlsin forests,compared to Based on an analysisof forest characteristicsat 100% for owlsin openings.Alternatively, edges and locations where owls responded to a vocal lure, openingscreated by clear-cutlogging may differ in Hayward and Garton (1988) concluded that the important waysfrom edgesand openingsin other Northern Pygmy-Owlwas a habitat generalist.At contexts.For example, transitionalvegetation was least two plausibleand non-mutuallyexclusive hy- all but absent in our study area. Also, our defini- pothesescan explain the differencesbetween their tion of a 30-m buffer to define edgeswas arbitrary, conclusionsand ours. First, owlsmight behavedif- and other definitions of edge might lead to differ- ferently on different study areas and second, owls ent conclusions.Lastly, we assumedthat the owls might use different criteria to select habitat for we tracked were foraging. While we attempted to home-range location, calling, and foraging. The strengthenthis assumptionby focusingon nesting owlswe studiedoccupied home rangesthat encom- males and eliminating locationsnear nests,we can- passeda diversearray of vegetationcategories, but not demonstrate that our data reflect use for for- they predominantlyused a subsetof thosecatego- aging. If owls used different vegetation categories ries.Additionally, calling locationsand foraginglo- for different activities,important use of some cat- cations appeared to differ (see below). Two im- egories might be obscuredby our analyses.These portant differencesbetween our studyand that of alternative explanations should be tested before Hayward and Garton are that they apparently sur- general conclusionsregarding edge associationsof veyedfor owlsat night, whereasour locationswere Northern Pygmy-Owlsare drawn. crepuscularand diurnal, and their studyincluded We also failed to find support for the idea that many vegetationtypes across a broad range of el- Northern Pygmy-Owlsuse nestsnear edges. How- evations. JUNE2003 PYGMY-OWLHABITAT USE 123

Our resultssuggest the need for care when eval- greater structural diversity.This suggeststhat log- uating habitat associationsof Northern Pygmy- ging practicesthat do not result in monocultural Owls fYom opportunistic sightingsor vocal lure sur- plantations may have a lesserimpact. veys. First, of 21 owls radio-marked, 11 were Our findings are similar to those of a recent initially detected respondingto playbacksfrom for- study of habitat use by the Eurasian Pygmy-Owl est edges (unpubl. data), yet our analyseson a sub- (Glaucidium passerinum) (Strom and Sonerud set of nine of these owls did not indicate dispro- 2001). Home range sizesfor males in their study portionately heavy use of edge habitat. This can (100% MCP, 40-600 ha) overlapped our estimates most easilybe explained if owlsreacted to calling and habitat use patterns were similar, except that surveysby moving toward the perceivedsource be- they found that Edgewas the highest ranked cate- fore vocalizing. Proudfoot et al. (2002) document- gory for the Eurasian Pygmy-Owl.However, they ed the movement of Ferruginous Pygmy-Owls defined Edgeto be a 10-m-wide strip, where forest (Glaucidiumbrasilianum) toward calling stations.In vegetation bordered open areas. To facilitate com- our case, because surveyswere conducted from parisons,we reanalyzed our data using their defi- logging roads, this would tend to pull owlstoward nition of Edge.It is important to note that although edges of clear-cuts.However, we cannot exclude Edgehad the highest relative rank in their study,it the possibilitythat edges are selected as sentinel did not differ significantlyfrom either of the next perch areasfor calling.Second, of the same21 ra- two highest ranked cover types (analogous to our dio-marked owls, only 12 establishedterritories in- StructurallyDiverse and Late StemExclusion vegeta- clusiveof their original responselocation (unpubl. tion categories). In our reanalysis,the Structurally data). Lastly, we occasionallydetected responses Diverse category again had the highest relative from multiple owlson the sameterritory at differ- rank, followed by Edge, Late StemExclusion, Early ent times during a season, again suggestingthat Stem Exclusion, and Open-Sapling (results not calling location and territory location may be de- shown). However, the relative rank of Edgewas not coupled. From our experience, the surestway to significantly different than any category ex- document Northern Pygmy-Owls in residence cept Open-Sapling.Thus, even though the relative would be to repeatedly detect unsolicited vocali- rank for Edgeincreased by one, there was still in- zations from the same area. sufficient evidence to conclude that Northern Pyg- None of the owls that we radio-marked were re- my-Owlsused edgesdisproportionately. located in subsequentyears. We did monitor nest- Our natural history observationssupport much ing owls in consecutiveyears at the same site in of the consensusknowledge regarding Northern five cases(LB-94/95, 95/96, 96/97, WC-96/97, and Pygmy-Owls.We observedsynchronous fledging in BC-96/97; Table 1). Additionally, there were three four instancesand the dates of fledging were more sites used in one season, but vacant in the follow- similar within years than between (Table 1). The ing season (UB-94, LC-95, and SK-96), for a total owls that we observed also exhibited a diverse diet, of eight siteswhere a residentmale wasnot found consistentwith previousreports (Earhart and John- at the same site in the following season.Possible son 1970, Snyderand Wiley 1976, Holt and Leroux explanationsinclude: (1) nest-sitefidelity was low, 1996). However, we have no information on rela- (2) mortality was high, (3) the study area was a tive prey availability, and owls may exhibit prey population sink, and (4) radio-markingnegatively preferencesthat would not be apparent in simple affected the owlswe worked with by either increas- tallies. ing their mortality or inducing them to find new territories. Information on the annual movements ACKNOWI,EDGMENTS of Northern Pygmy-Owlswould be useful in eval- These data were collected with the invaluable ass•s- uating these hypotheses. rance of ElissaArnheim, Dave Manson, and the Sequim Seven of the eight nestswe located were in late Audubon Club. We are grateful to J. Duncan, G. Hay- successionalforests. This finding, coupled with ward, T. K•atzner,G. Proudfoot, R. Reynolds, and two high use in the same typesof fbrest suggeststhat anonymousreviewers for insightful commentsthat great- the loss of late-successionalforest may negatively ly improved the final manuscript.This researchwas sup- ported by a grant from the WashingtonState Department affect Northern Pygmy-Owls.However, one nest of Natural Resources(FY97-021) and generousin-k•nd wasin a mid-aged stand that differed from most of support from the U.S. Forest Service Pacific Northwest the mid-aged standsin our study area by having Research Station. 124 GIESEAND FOPSMAN VOL. 37, NO. 2

LITERATURE CITED LINKIIART, B.D., R.T. REYNOI.DS, AND R.A. RYDER. 1998. •EBISCHER,N.J., P.A. ROBERTS,AND R.E. KENWARD.1993. Home range and habitat of breeding Flammulated Owls in Colorado. Wilson Bull. 110:342-351. Compositional analysisof habitat use fi:om animal ra- MARSH:UJ, D.B. (ED.). 1992. Sensitive of dio-trackingdata. Ecology74:1313-1325. Oregon. Oregon Department of Fish and Wildlife, •kMERICANORNITHOLOGISTS' UNION (AOU). 1983. Check- Portland, OR U.S.A. list of North American birds, 6th Ed. American Or- PARRY,B.T., H.J. VAUX,AND N. DENNIS.1983. Changing nithologists'Union, Washington,DC U.S.A. BENT, A.C. 1938. Life histories of North American birds conceptionsof sustained-yieldpolicy on the national forests.J. Forestry81:150-154. of prey. Part 2. U.S.Natl. Mus. Spec.Bull. 170:400-439. PROUDFOOT,G.A., S.L. BEASOM,F. Ct•VEZ-RAMIREZ,AND EARHA•T,C.M. AND N.K. JOHNSON.1970. Size dimor- phisin and food habits of North American owls. Con- J.L. MAYS.2002. Responsedistance of Ferruginous dor 72:251-264. Pygmy-Owlsto broadcastedconspecific calls. J. Raptor Res'. 36:170-175. HAYWARD,G.D. AND E.O. G^RTON.1988. Resource parti- tioning among forest owls in the River of No Return REYNOLDS,R.T. ANDB.D. LINKtIART.1990. Extra-paircop- Wilderness,Idaho. Oecologia75:253-265. ulation and extra-rangemovements in Flammulated Owls. Ornis Scand. 21:74-76. HENDERSON,J.A., D.H. PETER,R.D. LESHER,AND D.C. --, R.A. RYDER,AND B.D. LINI4H^RT. 1989. Small for- SH^W.1989. Forestedplant associationsof the Olym- est owls. Natl. Wildl. Fed. Sci. Tech. Ser. 12:134-143. pic National Forest. USDA Forest Service Ecol. Tech. Paper R6-ECOL-TP 001-88. SEAMAN,D.E., B. GRIFFITH, AND R.A. POWELL. 1998. KER- HOET, D.W. AND L.A. LEROUX. 1996. Diets of Northern NELHR: a programfor estimatinganimal home rang- es. Wildl. Soc. Bull. 26:95-100. Pygmy-Owlsand Northern Saw-whetOwls in west-cen- tral . Wilson Bull. 108:123-128. SNYDER,N.F.R. ANDJ.W. WILE'. 1976. Sexual size dimor- , R. KLINE,AND L. SULI.rWXN-HOLT.1990. A descrip- phisin in hawks and owls of North America. Ornithol. tion of "tufts" and concealingposture in Northern Monogz20:1-96. Pygmy-Owls.J. RaptorP•s. 24:59-63. STROM,H. AND G.A. SONERUD.2001. Home range and JOHNSGARD,P.A. 1988. North American owls, 1st Ed. habitat selectionin the PygmyOwl Glaucidiumpasse- SmithsonianInstitution, Washington, DC U.S.A. rinum. Ornis Fenn. 78:145-148. KtE,J.G.,J.A. BALDWIN,AND CJ. EvANs.1996. CALHOME: WEBB,B. 1982. Distribution and nestingrequirements of a program for estimatinganimal home ranges. Wildl. montane fbrest owls in Colorado. Colo. Field Ornithol Soc. Bull. 24:342-344. J. 16:76-81. LEB^N,F. 1994. Program RESELECT.University of Idaho, Moscow, ID U.S.A. Received12 April 2002; accepted26 December 2002