BREEDING ECOLOGY OF BARRED OWLS IN THE CENTRAL APPALACHIANS ABSTRACT- Eight pairs of breedingBarred Owls (Strix varia) in westernMaryland were studied. Nest site habitat was sampledand quantifiedusing a modificationof theJames and Shugart(1970) technique (see Titus and Mosher1981). Statisticalcomparison to 76 randomhabitat plots showed nest sites werb in moremature forest stands and closer to forest openings.There was no apparent association of nestsites with water. Cavity dimensions were compared statistically with 41 randomlyselected cavities. Except for cavityheight, there were no statistically significant differences between them. Smallmammals comprised 65.9% of the totalnumber of prey itemsrecorded, of which81.5% were members of the familiesCricetidae and Soricidae. Birds accounted for 14.6%of theprey items and crayfish and insects 19.5%. We also recordedan apparentinstance of juvenile cannibalism. Thirteennestlings were produced in 7 nests,averaging 1.9 young per nest.Only 2 of 5 nests,where the outcome was known,fledged young. The Barred Owl (Strix varia) is a common noc- STUDY AREA AND METHODS turnal raptor in forestsof the easternUnited States, The studywas conducted in Green Ridge State Forest (GRSF), though few detailedstudies of it havebeen pub- Allegany County, Maryland. It is within the Ridge and Valley lished.Most reportsare of singlenesting occurr- physiographicregion (Stoneand Matthews 1977), characterized by narrowmountain ridges oriented northeast to southwestsepa- encesand general observations(Bolles 1890; Carter rated by steepnarrow valleys(see Titus 1980). 1925; Henderson 1933; Robertson 1959; Brown About 74% of the countyand nearly all of GRSF is forested 1962; Caldwell 1972; Hamerstrom1973; Appel- Major foresttypes were describedby Brushet al. (1980).Predom- gate 1975; Soucy 1976; Bird and Wright 1977; inant tree speciesinclude white oak (Quercusalba), red oak (Q. Leder and Walters 1980). Habitat was described rubra),chestnut oak (Q.prinus), scarlet oak (Q. coccinea),red maple (Acerrubrum), and pignut and mockernut hickories(Carya glabra qualitativelyby Nicholls and Warner (1972) and and C. tomentosa).Predominant understory species include flow- Fuller (1979). Barred Owl food habits were re- ering dogwood(Cornus florida), sassafras(Sassafras albidum), ser- ported by Cahn and Kemp (1930), Errington viceberry(Amelanchier spp.), and saplingsof the dominanttrees. (1932), Errington and McDonald (1937), Wilson The studyarea wassystematically searched for activenests from late Februarythrough May in 1981and 1982.During 1982,tape (1938), Mendall (1944), Hamerstrom and recorded Barred Owl calls were broadcast in order to elicit re- Hamerstrom (1951), Blakemore (1960) LeDuc sponsesand help localizenesting pairs. (1970),and Korschgenand Stuart(1972). The food Nest siteswere plottedon 7.5 rain USGStopographic maps and habits studied, however, were all from midwestern County Soil ConservationService maps. A nestsite was defined as states,except Mendali's (1944) study from Maine. a 0.4 ha plot (11.3 m radius) centered on the nest tree. This size plot was consideredmore time and field efficient than either Dunstanand Sample(1972) reported the number smalleror largersize plots when making quantitative estimates of of fledglingsfrom 1 cavityeach year for 5 years,but the vegetation(Lindsey et al. 1958,James and Shugart1970). provided no other productivity information. Nestswere checkedperiodically each season to obtainnesting Clutch sizesin variousgeographic regions can be chronologyand productivity information. At the same time, re- gurgitatedpellets found in the cavitieswere collected and anyprey found in Bent (1961) and Murray (1976). remains were noted. This study was conducted in an area where 4 At the end of the nestingseason, vegetation at eachactive nest diurnal raptor species,the Red-shoulderedHawk sitewas sampled using a modificatonof the Jamesand Shugart (Buteolineatus ), Broad-wingedHawk (B.platypterus ), technique (1970), as describedby Titus and Mosher (1981). Thirty-four variableswere measuredor derived at eachsite (Table Red-tailed Hawk (B. jamaicensis)and Cooper's 1). The type of cavityin whicha pair nested(hollow tree stub,hole Hawk (Accipitercooperi), were alsounder study(see from disease,excavated hole, or hole from broken limb) and Titus and Mosher 1981,Janik and Mosher 1982). successionalstage of the cavitytree (Fig. 1) were recorded. Our objectiveswere to quantitativelydescribe veg- Height to cavityentrance was measured with a meter tape for etation structure at Barred Owl nest sites and com- treesclimbed, otherwise height measurementsand percentslope weremeasured with a Hagaaltimeter. Percent canopy, understory pare it with surrounding habitat, measure and and ground coverswere basedon 40 ocular tube readings, 10 comparedimensions of cavitiesused by them with alongeach of 4 transectsstarting at the nesttree and extendingin those from randomly selectedcavities, describe eachof the cardinal compassdirections. their food habitsfor this geographicregion, and We comparednest site data with random habitat samplescol- lected by Titus and Mosher (1981) to determine if vegetation determinetheir breedingchronology and produc- structure around nest trees differed from surrounding habitat. tivity. Variablesmeasured at random plotsare listedin Table 1 except 49 RAPTORRESEARCH 18(2):49-58 50 DEVERE^UX AND MOSHER VOL. 18, NO. 2 Table 1. Quantativehabitat variables and cavitycharacteristics used in analysisof Barred Owl nest site habitat 1. ALTITUDE Altitude of plot in meters;taken from U.S.G.S.7.5-min. quadrangles 2. SOIL Soil-woodssuitability; measures suitability for tree productivity;class 1 indicateshigh produc- tivityand class6 indicateslow productivity(Stone and Matthews1977) 3. SITINDX Siteindex; based on SOIL and the tree speciespresent in the plot (Stoneand Matthews1977) 4. WATER Distance to water in meters 5. DISFOROP Distanceto the nearestforest opening in meters;measured to the nearestbreak in forest continuity,such as created by trail, road, field, etc. 6. PERSLOP Percentslope of plot 7. CANHT Canopyheight of the plotin meters;the meanof 5 measurementstaken to the top of the canopy 8. CANEVER Percentageevergreen canopy cover 9. CANTOT Percentagetotal canopy.cover 10. UNDEVER Percentageevergreen understorycover 11. UNDTOT Percentagetotal understorycover 12. GRNDEVER Percentageevergreen ground cover 13. GRNDTOT Percentagetotal ground cover 14. SHRUBDEN Shrubdensity (James and Shugart1970, James 1978) 15. SHRUBIND Shrub index (Titus 1980) 16. NOSPTREE Number of speciesof overstorytrees in the plot 17. NOSPSHRB Numberof speciesof shrubsand saplingsin the plot 18. NOTREES Number of overstorytrees in the plot 19. UND14 Numberof understorystems 1-4 cm diameterin the plot 20. UND58 Number of understorystems 5-8 cm diameterin the plot 21. UNDGT8 Numberof understorystems greater than 8 cm diameterin the plot 22. DBHLT26 Number of overstorytrees less than 26 cm dbh in the plot 23. DBH2650 Number of overstorytrees 26-50 cm dbh in the plot 24. DBHGT50 Number of overstorytrees greater than 50 cm dbh in the plot 25. BASAL Basalarea in m2/hatbr overstorytrees 26. DBH* Diameter at breastheight of nest tree 27. TREEHT* Height of cavitytree in meters 28. CAVHT* Height to lowestpoint of cavityentrance in meters 29. %CAVHT* Percentagecavity height; calculated as: (GAVHT/GANHT) (100) = %GAVHT 30. TREEDIAM* Diameter of cavitytree at cavityheight 31. HORIZONT* Horizontal length of cavityopening in cm 32. VERTICAL* Vertical length of cavityopening in cm (Table 1 continued) SUMMER 1984 APPALACHIAN BARRED OWLS 51 (Table 1 concluded) 33. CAVDIAM* Insidediameter of cavityin cm;measured from insideof entranceto backwall; for hollowtree stubs,the largestdiameter is recorded 34. CAVDEPTH* Cavitydepth in cm; measuredfrom lowestpoint of cavityentrance to baseof cavity. (* = variablesunique to cavitiesand cavity trees). for the cavityand cavitytree specificvariables. (Siegal 1956) tested for similarity betweennest site habitat and Dimensionsof 41 randomlyselected, unoccupied cavities were random habitatplots, and nest site cavityand random cavityd•- measuredand compared with nest cavities to providea measureof mensions.Spearman rank correlation coeffidents (Siegal 1956) cavitysizes availble to BarredOwls and assesscavity selection. The were calculatedto determine the extent of correlation among random samplingof cavitieswas stratified. Transects, approxi- structural features of habitat and among cavity characteristics mately100 m apart, 1.6km longextending on bothsides of a road X 2 goodness-of-fittests were usedon pooledsamples of nestsite running the lengthof the studyarea, were randomlychosen. A and randomcavities to determineif differencesexisted among the coin flip determined which side of the road the transectwas numberof eachcavity type found and numberof cavitytrees •n walked. Every third cavityencountered was measuredbut no eachsuccessional stage. Test results were considered significant if morethan 3/transectto avoidmeasuring too many within a single P < 0.05. habitattype. The criteriaibr acceptinga randomcavity was that it be at least2 m from the groundand haveat leasta 15cm diameter RESULTS AND DISCUSSION opening,or, fbr a hollowtree stub,a 25 cm dbh. Minimum sample sizeswere calculatedfbr each variable to Habitat. -- Eight-Barred Owl nestswere located. determineif random samplingwas adequate. Sample sizes were The 4 found in 1981 were not reused in 1982. Nest consideredadequate if they met the criteria of remaining