Roosting Behavior of Postfledging Eastern Screech-Owls

ROOSTING BEHAVIOR OF POSTFLEDGING EASTERN SCREECH-OWLS

JAMESR. BELTHOFF1 AND GARY RITCHISON Departmentof BiologicalSciences, Eastern Kentucky University, Richmond,Kentucky 40475 USA

AI•S?RACT.--Weexamined roosting behavior of adult and juvenile EasternScreech-Owls (Otusasio) during the postfledgingperiod using radiotelemetry.We located1,107 screechowl roostsites in 39 speciesof trees,shrubs, and vines. Nearly half (47.8%)were in eastern red cedar(Juniperus virginiana), shagbark hickory (Caryaovata), black locust(Robinia pseudo- acacia),and black walnut (Juglansnigra). Owls usedopen limb roosts(46.4%), tangle roosts (36.2%),and conifer roosts(17.4%). Tree cavitieswere usedrarely (n = 3). Open limb roosts were most commonearly in the postfledgingperiod, whereas use of tanglesfor roosting increasedlater in the season.The mean roostsite was 10.2 + 0.2 m high in a tree 14.2 + 0.2 m tall with a DBH (diameterat breastheight) of 23.8 + 0.4 cm. Basedon a random sample of 800 potentialroost trees, screech-owls used significantly shorter trees than thoseavailable. However, we found no significantdifference in meanDBH betweenused and availabletrees. On average,screech-owls roosted 252 + 5.3m from their nestsand moved64 + 3.5 m between daily roostsites. Juvenile and adult owls differed little in selectionof roostsites, although juveniles used a greater variety of trees.Paired adults did not differ in roost-siteuse. The meandistance between roost sites of young owls and their parents(both male and female) increasedsignificantly after the fifth week postfledging.After this period, juvenile owls roostedunaccompanied by adults much more often, which suggestsyoung gain somein- dependencefrom adults.We noted significantdifferences among families for all roost-site variables,with mostvariation explained by differencesin areasoccupied by families.Entire familiesroosted together in the sametree 31 + 7.5%of the time (range:16.7-51.5%). Coef- ficientsof associationat roostsites between adults and their young were similar for both membersof a pair. We suggestthat adult EasternScreech-Owls do not divide their broods. Received26 June1989, accepted 8 February1990.

EASTERNScreech-Owls (Otus asio) are small, juvenile screech-owlsand their parentsduring nocturnal birds found throughout much of the the time between fledging of the young and easternUnited States.Many aspectsof screech- their subsequentdispersal. owl biology have been examined(e.g. Hrubant Our objectivewas to examine the roosting 1955; Owen 1963a, b; Marshall 1967; Ross 1969; behavior of radio-taggedadult and juvenile VanCamp and Henny 1975; van der Weyden EasternScreech-Owls during the postfledging 1975; Turner and Dimmick 1981; Smith and Gil- period in central Kentucky. We sought to de- bert 1984; Cavanagh and Ritchison 1987; Heg- termine if screech-owls selected roost trees non- dal and Colvin 1988; Ritchison et al. 1988; Belt- randomlyfrom availabletrees, if roostsof adults hoff and Ritchison 1989, MS), but little is known and their young differed, if roostingbehavior about the roostingbehavior of EasternScreech- varied either among families or over time, and Owls. Although previous work suggeststhat if parentsdivided their broodsinto subgroups. screech-owlsusually roost in tree cavities,most observationswere made during the autumn, winter, and early spring (Merson et al. 1983, STUDY AREA AND METHODS Smith et al. 1987). Few data are available on the roosting of screech-owlsduring summer.Even We examined the postfledgingbehavior of Eastern less is known about the roosting behavior of Screech-Owlsbetween mid-May and lateJuly in 1985 and 1986 at the 680-haCentral KentuckyWildlife Management Area, located 17 km southeastof Rich- mond, MadisonCounty, Kentucky.This area com- • Presentaddress: Department of BiologicalSci- prises small deciduous woodlots and thickets inter- ences,Clemson University, Clemson,South Carolina spersedwith cultivated fields and old fields. 29634 USA. We capturedadult EasternScreech-Owls directly 567 The Auk 107:567-579. July 1990 568 BELTHO• AND RITCHISOIq [Auk, Vol. 107

from nest boxes and natural tree cavities, or with mist In 1985 we radio-taggedand monitored the roost- nets while using their young or playbackof bounce ing behavior of all individuals in three families (2 songs(Ritchison et al. 1988) as lures. Juvenile owls adults and 3 young in each; referred to as Off-prop- were capturedfrom nest cavitiesjust before fledging erty, Muddy Creek, and Trap Range families). All (i.e. leaving the nest cavitypermanently). In the case individuals (4 adults and 5 young) in two families of one family, we capturedyoung from tree limbs just (Stream and Hilltop families) were radio-tagged in after they fledged. We fitted capturedadult and ju- 1986. Three days before the young fledged, a Great venile screech-owls with radio transmitters (Wildlife Horned Owl (Bubovirginianus) killed the adult female Materials Inc., Carbondale, Illinois) backpack style in the Hilltop family. We capturedand radio-tagged with woven nylon cord (Smith and Gilbert 1981).The the adult male and three young from another family transmitter and harnessweighed <8 g. (GoosePen) in 1986.We were unable to radio-tagthe After young owls fledged,we locateddiurnal roost GoosePen adult female, who was observed only oc- sites of adults and juveniles at least four times per casionally.We also monitored a seventh family, the week in 1985 and daily in 1986 until young dispersed 1986 Trap Range family. We were able to radio-tag from natal territories (55 _+ 1.3 days after fledging; only 3 of 4 young in this brood, consideredthese data Belthoff and Ritchison 1989). We located owls be- supplemental,and excludedthem from analyses.All tween noon and 1800 with portable receivers(Wild- referencesto the Trap Range family refer to 1985 life Materials TRX-24 or Telonics TR-2) and hand-held unless otherwise noted. 2-element yagi antennas (Telonics Inc., Mesa, Ari- Data analysis.--We used analysis of variance (AN- zona). Owls occasionally(<5% of observations)flew OVA) to testfor effectsof age,family, and time (weeks) from roostsas we approached,particularly near the on the roostingbehavior of EasternScreech-Owls. If beginningof the postfledgingperiod. No more than significanteffects existed, then we estimatedmeans one family member ever flew from the roost,and we (least squaresmeans) for ages,families, and time and were usually closeenough to determine the original conductedpairwise t-tests to make comparisonsof roost location. The roost sites selectedby owls after thesemeans. To examinepossible effects of increasing flushing were not included in the analyses.Although ageon the roostingbehavior of juvenile screech-owls we did not record information about flushing behav- and their parents,we pooled observationsinto week- ior, adultswere more likely to fly than young. Owls ly categories.Most young EasternScreech-Owls dis- initiated fewer flight-intention movements and flew persedduring the eighth and ninth weeksafter fledg- from roostsless frequently as the study progressed, ing, which resultedin someempty cellsin the ANOVA perhapsbecause the owls habituatedto our presence. for week 9. We combined data from these 2 weeks for We categorizedeach roost as either cavity, open analyses.To avoid possible temporal pseudoreplica- limb, conifer, tangle (Merson et al. 1983), or ground. tion (Hurlbert 1984), we consideredthe effect of time If an owl roostedamong vines in either a deciduous as a repeated measure in the ANOVA (Krebs 1989: or coniferoustree, we classifiedthe roost as a tangle. 274). For these analyseswe consideredroost sitesof For each roost site, we recorded tree species,roost individualsas independentobservations. Thus, if all height, tree height, percentheight (roostheight/tree five membersof a family roostedin the sametree on height), diameterat breastheight (DBH), distancefrom the same day, or if one owl roosted in the same site nest, and distance from previous roost (only those on five different days, we consideredthese as five distancesrecorded on consecutivedays were included independentroosts. We assumedthat sitesused more in the analyses).We alsonoted which owls in a family than once were favorable. In this manner, character- unit roostedtogether (in the sametree). We calculated isticsof roostsused more than once,or by morethan coefficientsof associationfrom the formula: 2ab/(a + one family member, were weighted proportional to b), where a is the total number of times owl A was their use. observed, b is the total number of times owl B was We used Mann-Whitney U-tests (Zar 1974) to ex- observed, and ab is the total number of times owls A amine differencesin the numberof roost-treespecies and B were observedtogether (Cole 1949).When fam- usedby owls of different age classes,to examine dif- ily members roosted separately, we measured dis- ferences between sexes of adults in coefficients of tancesbetween individuals (usingaerial photographs associationwith offspring,and to examinedifferences of the study area to estimatedistances >500 m). We in trees used for roosting and available roost trees. estimatedtree and roost heights with a clinometer. We conducted Chi-square goodness-of-fittests to ex- Finally, we usedpoint-quarter sampling (Greig-Smith amine differencesamong frequencies of observedand 1964) to determine the relative abundance of trees expectedroost species.If a cell in any contingency available for roosting.We randomly positionedtran- table had an expectedvalue < 1, specieswere pooled sectsthrough five woodlotsused by owls, and sam- to raisethe value and reducethe degreesof freedom. pled 40 stations (i.e. 160 trees) along each transect. All testswere two-tailed (nondirectional),and signif- We estimated heights and DBHs of available roost icance levels were set at 0.05. Values are means and trees using these same trees. standard errors (œ + SE). July1990] EasternScreech-Owl Roosts 569

RESULTS gle roostwas 8.8 + 0.3 m (n = 393) high in a tree with a mean height of 11.9 + 0.3 m (n = ROOST-SITE CHARACTERISTICS 399) and a mean DBH of 22.1 _+ 0.7 cm (n = 396). Tangle roostswere located in trees sig- OveralL--We located 1,107 roost sites of adults nificantlysmaller than availabletrees, both in (n = 10) and juveniles (n = 17) during postfledg- height (Mann-WhitneyU-test, P < 0.0001)and ing in 1985 and 1986. Screech-owlsused 39 DBH (Mann-Whitney U-test,P < 0.0002).Owls speciesof trees, shrubs, and vines, plus un- roosted in tangles in 31 speciesof trees and identified snagsfor roosting (Table 1). Nearly shrubs, with black locust (Robiniapseudoacacia, half (47.8%) of all roosts were located in four 17.6%), eastern red cedar (Juniperusvirginiana, speciesof trees. Speciesin owl territories but 9.3%), and box elder (Acer negundo,6.8%) used not usedinclude pawpaw (Asiminatriloba), sugar most frequently. The mean conifer roost was maple (Acer saccharum),tulip poplar (Lirioden- 6.9 + 0.2 m (n = 184) high in treeswith a mean drontulipifera), and white mulberry (Morusalba ). height of 9.9 + 0.3 m (n = 192) and mean DBH The mean roost site was 10.2 + 0.2 m high of 20.7 + 0.9 cm (n = 192). Conifer roosts were (range:0-27.1 m, n = 1,068)in a tree 14.2 + 0.2 alsoin treessignificantly smaller both in height m tall (range:0.9-36.0 m, n = 1,103)with a DBH and DBH (Mann-Whitney U-test, P < 0.0001 of 23.8 + 0.4 cm (range:2.5-89.4 cm, n = 1,094). for each)than available trees.Eastern red cedar The mean height of a sample of 800 trees was was the only speciesof conifer used. 15.2 + 0.3 m, significantly taller than trees used Variationamong families.--All six families by roosting owls (Mann-Whitney U-test, P < roostedin tree speciesat frequenciesdifferent 0.0001). The mean DBH of available trees was from expectedbased on availability (Table 1). 24.9 + 0.5 cm, but this did not differ signifi- Threefamilies (Muddy Creek, Off-property, and cantly from DBH of treesused by roostingowls GoosePen) roostedin easternred cedarsignif- (Mann-Whitney U-test, P > 0.140). Owls roost- icantly more than expected.Other tree species ed an averageof 252.5 + 5.3 m (range: 1.8-691.3 used more often than expectedby individual m, n = 1,107) from their nest sites. Although families included shagbark hickory (used by roostsites used by owls on successivedays were Trap Range),southern red oak (Quercusfalcata) 63.7 + 3.5 m (range: 0-740.7 m, n = 861) apart, and black gum (Nyssasylvatica) (used by Off- individuals used the same roost on consecutive property),black locustand black walnut (used days 10.7% of the time. Juvenile owls roosted by Stream),and apple (Malussp.) (usedby Hill- 53.4 + 4.5 m (range: 0-717.8 m, n = 734) from top). Tree speciesused lessoften than expected the adult male, and 36.0 + 3.8 m (range: 0-512.1 basedon availabilityincluded sassafras (Sassaf- m, n = 522) from the adult female. rasalbidum), red maple (Acerrubrum), and south- Eastern Screech-Owls used open limb roosts ern red oak (all usedby Trap Range);green ash most frequently (46.4%), then tangle roosts (Fraxinuspennsylvanica) (used by Off-property (36.2%) and conifer roosts (17.4%). Only one and GoosePen), shagbarkhickory, black gum, individual (1986 Trap Range adult female) and sweet gum (all used by Muddy Creek), roosted in tree cavities (n = 3) during the post- chinquipin oak (Quercusprinoides) and Shumard fledging period, all during the first week after oak (Q. shumardii)(used by Stream), and white the young fledged. Once, a young owl roosted ash(Fraxinus americana) (used by Hilltop). Three on the ground beneatha fallen branch. families(Off-property, Stream, and GoosePen) The mean open limb roost was 12.7 + 0.3 m usedsnags less often than expected. (n = 491) high in a tree with a mean height of All six owl families roosted in trees that dif- 17.6 _+0.4 m (n = 510) and a mean DBH of 26.2 feredsignificantly in size(height and DBH) from + 0.6 cm (n = 504). Open limb roostswere sig- available trees. Four families roosted in trees nificantly higher and in treessignificantly taller than either tangle or conifer roosts(ANOVA; smaller than those available, while the remain- P < 0.05). Owls roosted on open limbs in 34 ing two families roostedin larger than expected speciesof trees and shrubs.Shagbark hickory trees (Fig. la, b). We noted many significant (Caryaovata, 25.4%), black walnut (Juglansnigra, differencesin mean roost height, tree height, 7.8%), and sweet gum (Liquidambarstyracifiua, DBH, percent height, distance from nest, and 7.8%)were usedmost frequently. The mean tan- distancebetween successiveroosts among faro- 570 BELTHOFFAIqD RITCHISOIq [Auk,Vol. 107

T^I•LE1. Percentagetree and shrub speciesselected by six families of EasternScreech-Owls (n = 1,107) during the postfledgingperiod in central Kentucky.We determinedpercentages of availabletree species (Avail. %) usingpoint-quarter sampling of 160 randomlyselected trees along transects through eachowl territory.The Muddy Creekand Streamfamilies used the samewoodlot, and only one censusof 160 trees was conductedfor both. Samplesizes are in parenthesesbeneath each age class.

Family Off-property Muddy Creek Trap Range Ad. Juv. Avail. Ad. Juv. Avail. Ad. Juv. Avail. Species (67) (104) % (73) (110) % (65) (95) % Juniperusvirginiana 17.9 25.0 7.5 56.1 48.2 6.9 3.1 4.2 0.6 Carya ovata 0.0 1.0 0.0 0.0 0.0 10.0 53.8 61.1 31.3 Robiniapseudoacacia 20.9 13.5 23.1 1.4 0.0 0.0 0.0 0.0 0.0 Juglansnigra 3.0 3.8 1.9 1.4 0.0 3.8 0.0 9.4 0.6 Liquidambarstyracifiua 1.5 1.0 0.0 2.7 1.8 10.0 15.4 9.4 10.0 Cercis canadensis 0.0 0.0 0.6 0.0 1.0 5.6 0.0 0.0 0.0 Acer negundo 5.9 3.8 0.0 1.4 2.7 3.1 0.0 0.0 0.0 Ulmus americana 4.5 3.8 11.9 9.6 7.3 2.5 0.0 0.0 2.5 Quercusfalcata 16.4 21.2 0.0 0.0 0.0 0.0 1.5 3.2 12.5 Fraxinus americana 1.5 0.0 1.3 5.5 9.1 6.9 0.0 0.0 1.3 Morus rubra 9.0 5.8 0.6 5.5 3.6 0.6 0.0 0.0 1.3 Celtis occidentalis 4.5 3.8 6.3 2.7 1.8 2.5 0.0 0.0 0.0 Ostryavirginiana 0.0 0.0 0.0 0.0 0.0 3.8 0.0 0.0 0.6 Malus sp. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Quercus shumardii 0.0 0.0 0.0 0.0 0.0 4.4 12.3 6.3 1.3 Cornusflorida 0.0 3.8 0.0 0.0 2.7 1.3 0.0 0.0 0.0 Gleditsia triacanthos 0.0 1.0 2.5 4.1 0.0 1.3 0.0 0.0 0.0 Sassafrasalbidum 0.0 0.0 0.0 0.0 0.0 3.8 0.0 0.0 7.5 Nyssasylvatica 5.9 5.8 0.0 0.0 0.0 5.6 0.0 0.0 0.0 Diospyrosvirginiana 1.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Quercus alba 0.0 0.0 0.0 0.0 0.0 1.3 1.5 0.0 1.9 Q. imbricaria 4.5 1.0 0.6 2.7 2.7 1.9 1.5 0.0 3.1 Platanus occidentalis 0.0 1.9 8.8 0.0 0.0 1.3 0.0 1.1 1.3 Others a 3.0 3.8 35.0 6.9 19.1 23.7 10.8 5.3 29.8 x2-value, df 142.0, 18 158.6, 23 87.4, 11 P-value 0.0001 0.0001 0.0001

ßOther speciesused for roosting(Overall [%] used,Overall [%]sampled): Carya tomentosa (0.8, 0.4); Quercusstellata (0.6, 1.3);Carya laciniosa (0.6, 0.9); Rhamnuscaroliniana (0.5, 0.0); Quercusprinoides (0.5, 1.3); Acer rubrum(0.5, 2.0); Allanthusaltissima (0.5, 0.0); Quercusvelutina (0.5, 0.5); Ulmus rubra(0.5, 0.5); Caryacordiformis (0.4, 1.4);Fraxinus pennsylvanica (0.4, 6.9); Prunusserotina (0.3, 0.1); Tilia americana(02, 0.0); Vitisspp. (0.1, 0.0); Loniceraspp. (0.1, 0.0);Carya glabra (0.1, 0.1);unidentified snag (2.5, 7.6); utility pole/vines(0.3, 0.0);ground (0.1,0.0).

ilies (Table 2). Families also differed signifi- veniles in the Stream family used open limb cantly in the mean distancebetween roostsites roostsmore, and tangle roostsless, than their of adults and their young (Table 2) and in the parents (Chi-square test, P < 0.05), whereas roosttypes they used(X 2 = 249.6, df = 10, P < Goose Pen juveniles used conifer roosts less 0.001; Fig. 2). often, and tangle roostsmore often than their Adult malesvs. females,and adults vs. juve- parents (P < 0.001). niles.--Overall, we observedno significantdif- We noted no differences between adult and ferencesin the type or characteristicsof roost juvenile owls in roostheight, roost-treeheight, sites used by adult male and female screech DBtt, percent height, or the distance of roost owls (Table 3); and there was no family effect. sites from nest trees (Table 3). The mean dis- Adult and juvenile owls did not differ signifi- tancebetween daily roostsites was significantly cantly in their use of roost types (Table 3), al- greater for adults than juveniles for all six fam- though we noted a significantfamily effect (P ilies pooled (P < 0.05), but we also noted a < 0.01).Significant differences existed between significant family effect (P < 0.0001). Within adults and juveniles in two families only. Ju- individual families, distances between daily July 1990] Eastern Screech-Owl Roosts 571

TABLE 1. Extended.

Family Stream Goose Pen Hilltop Overall Ad. Juv. Avail. Ad. Juv. Avail. Ad. Juv. Avail. Avail. (108) (107) % (66) (155) % (54) (101) % Total % 13.0 9.3 6.9 48.5 22.6 4.4 0.0 1.0 0.0 20.8 3.9 21.3 18.7 i0.0 1.5 2.6 1.3 0.0 0.0 5.6 12.9 9.6 12.0 0.0 0.0 9.1 23.2 23.8 5.6 5.9 3.1 8.4 10.0 14.8 13.1 3.8 0.0 0.0 3.8 13.0 9.9 8.1 5.7 3.6 6.5 8.4 i0.0 0.0 0.0 0.6 7.4 5.9 18.8 4.6 7.9 0.9 1.9 5.6 7.6 23.9 16.9 0.0 3.0 0.6 4.4 4.8 6.5 5.6 3.1 0.0 0.0 0.0 14.8 13.9 4.4 4.3 1.5 0.9 0.0 2.5 3.0 3.9 6.9 9.3 6.9 11.9 3.9 7.1 0.0 0.0 0.0 0.0 0.0 0.0 1.8 0.0 0.0 3.4 2.5 3.7 5.6 6.9 1.5 1.9 0.6 0.0 0.0 9.4 2.6 3.9 0.0 0.0 0.6 4.6 2.6 0.6 0.0 0.0 0.0 2.4 0.9 0.0 1.9 2.5 0.0 1.3 3.8 0.0 9.9 4.4 2.3 3.4 7.4 11.2 3.8 0.0 0.0 0.0 0.0 0.0 3.8 1.8 1.6 0.0 0.0 0.0 0.0 0.0 0.0 18.5 8.9 0.0 1.7 0.0 0.0 0.0 4.4 0.0 0.0 0.0 0.0 4.0 2.5 1.6 1.6 0.0 5.6 1.3 1.5 1.3 0.6 0.0 1.0 0.0 1.5 0.4 0.9 0.0 1.3 0.0 1.9 3.1 7.4 3.0 2.5 1.4 1.9 0.9 2.8 3.8 4.6 5.2 2.5 0.0 0.0 1.3 1.4 3.0 0.9 1.9 5.6 0.0 0.0 0.6 0.0 0.0 0.0 1.2 1.4 0.0 0.0 0.0 7.6 3.9 6.3 0.0 0.0 0.0 i.i 1.3 5.6 3.7 1.3 0.0 0.0 0.0 1.8 0.0 5.0 i.i 1.6 0.0 0.0 1.9 0.0 0.0 0.0 0.0 1.0 0.6 1.0 1.3 0.0 1.9 1.3 3.0 0.0 3.1 3.7 2.0 2.5 1.0 3.4

4.5 8.4 23.1 7.5 5.7 21.2 16.4 23.0 13.7 9.4 23.0 93.1,23 85.1,15 81.8,19 0.0001 0.0001 0.0001

roostsites were significantlydifferent for adults both adults (Table 4). There were, however, sig- and juveniles (P < 0.01) in the Hilltop family nificant interactions(P < 0.001) with family for only (œ= 177.5 + 37.5, n = 51 and ;?= 56.3 + roost height, tree height, DBH, and percent 10.4, n = 98 for adults and juveniles, respec- height. In two families (Muddy Creek and tively). Stream),unaccompanied young roostedin sig- Juvenile owls used significantly more tree nificantlysmaller trees (height and DBH) than speciesfor roostingthan adults(Mann-Whitney accompaniedyoung. Young in thesetwo fam- U-test, P < 0.02). Adults used 14 + 0.9 species ilies also roosted lower when not accompanied for roosting (range: 11-17, n = 6 families), and by adults (both in actual height and percent juvenile owls used 16 + 1.9 species(range: 8- height). Hilltop family young roosted propor- 22). We noted few differencesin the roostspecies tionately lower (percent height) when unac- usedby adult and juvenile owlswithin families, companiedby adults. Neither the distancebe- although occasionallarge differences (greater tweendaily roostsites nor the distanceof roost than approximately10% difference in use) ap- sitesfrom nestswas influenced by the presence peared (Table 1). or absenceof adults (P > 0.860 for each).Finally, Accompaniedvs. unaccompaniedyoung.--Un- unaccompaniedand accompaniedyoung did not accompaniedyoung roosted in significantlydif- differ in use of roosttypes (x2 = 3.71, df = 2, P ferent sites than young roosting with one or > 0.16, n = 5 families). 572 BELTHOFFAND RrrcHISON [Auk,Vol. 107

100.

*** [] Observed [] OFF-PROPERTY 30] [] MUDDY CREEK 80 - [] TRAPRANGE [] S•.EAM [] GOOSEPEN

GP TR STR MC OP ! Ir 0 OpEN LIMB CONIFER TANGLE Roost Type [] Observed Fig. 2. Frequency(%) of open limb, tangle, and [] Expected coniferroosts used by six familiesof EasternScreech-

NS Owls (adultsand iuvenilespooled) during the postfledging period in central Kentucky.Sample sizes: Off-property(173), Muddy Creek(183), Trap Range (161),Stream (214), Goose Pen (220), and Hilltop (155).

adult males and femalesdid not exhibit pref- erencesfor roostingwith particularyoung (Fig. GP TR STR MC OP 3). Family Fig. 1. A comparisonof (a) observedvs. expected TEMPORAL VARIATION tree heights and (b) observedrs. expectedDBHs. We generated expected values based on point-quarter Characteristicsof roostsites.--Eastern Screech- sampling of 160 potential roosttrees in each owl ter- Owls exhibitedsignificant variation in the types ritory and comparedthem with observedsites using of roostsites used as the postfledgingperiod a Mann-Whitney U-test (* = P < 0.05; ** = P < 0.01; progressed(X 2 = 244.8,df = 14, P < 0.001;Fig. *** = P < 0.001; NS = P > 0.05). Families: GP = Goose 4). During the first 2 weeks postfledging,owls Pen; TR = Trap Range;STR = Stream;MC = Muddy usedopen limb roostsmost commonly, and used Creek;OP = Off-property;HT = Hilltop. few tangle roosts.Together, these weeks ac- countedfor 50.9%of overall Chi-squarevaria- ASSOCIATION OF FAMILY MEMBERS tion. During week 4 after fledging,owls used each roost type equally. From week 5 after Entire families (n = 4) roostedtogether 31 _+ fledging until dispersalof young, owls used 7.5%of the time (range:16.7-51.5%). Although tangle roost sites more often than either open all membersof a family never roostedtogether limb or conifer roost sites. on the first day after fledging, they did so fre- Most roost-sitecharacteristics varied signifi- quently from the 2nd through the 44th day. cantly during the period from fledging until Between the 44th day and juvenile dispersal, dispersal.Because a significantfamily effect entire families no longer roosted together. (ANOVA) existed, we could not pool family Young screech-owlsroosted with adults (one or data. Mean roost height varied over time in all both) 63 + 7.7% of the time (range: 54-86%). six families, but we observed no consistent ten- Adult females (coefficient of association= 57 _+ dencies (Fig. 5a). We observed similar results 8.1;range: 42-81) roostedwith youngmore often for tree height, DBH, and percent height. In than adult males(43 + 7.1; range: 24-59), but contrast, the mean distance of roost sites from the differencewas not significant(Mann-Whit- the nestwas significantlylower during week 1 ney U-test, P > 0.20). Within families (n = 4), after fledging in all six families (Fig. 5b). The July1990] EasternScreech-Owl Roosts 573

TAI•Lœ2. Characteristics(LS œ+ SE)of roostsites used by six familiesof EasternScreech-Owls (adults and juvenilespooled). Means with the sameletter are not significantlydifferent (c•= 0.05). Samplesizes are in parentheses.

Roost Tree Percent Dist. from Dist. from Dist. from' Dist. from' height height height DBH previousroost nest adult female adult male Family (m) (m) (%) (cm) (m) (m) (m) (m)

Off- 9.7 _+ 0.3 B 13.6 _+ 0.5 B 72.2 _+ 1.2 B 25.8 _+ 1.0 B 68.2 _+ 10.2 BC 397.3 _+ 9.1A 43.9 _+ 12.1BC 81.9 _+ 8.0 A property (172) (172) (172) (172) (98) (173) (104) (101) Muddy 10.5 _+0.3B 14.2 _+0.5B 73.0 _+1.1B 25.0 _+0.9 B 30.4 _+10.2 D 102.5_+ 9.8 D 13.7 _+11.9 C 12.8 _+7.8C Creek (183) (182) (182) (183) (99) (183) (107) (107) Trap 20.6 _+0.4 A 24.4 _+0.5 A 83.3 _+1.2 A 33.6 _+1.0 A 21.0 _+10.4 D 290.7 -+ 9.4 B 57.9 -+ 12.7AB 24.9 _+8.3 BC Range (160) (168) (160) (160) (95) (161) (95) (94) Stream 10.4 _+ 0.3 B 15.5 _+ 0.4 C 71.8 _+ 1.1 B 23.4 ñ 0.9 B 56.3 -+ 6.9 C 103.2 _+ 8.1 D 44.0 _+ 11.9 BC 38.6 _+ 7.8 B (194) (214) (193) (211) (213) (215) (107) (107) Goose Pen 6.3 _+ 0.3C 9.8 _+ 0.4D 72.6 _+ 1.1B 16.6 _+ 0.9 D 80.1 _+ 7.0AB 405.9 -+ 8.0 A 74.2 -+ 9.9 AB 5.3 -+ 11.6 b (208) (223) (207) (219) (207) (222) (154) (48) Hilltop 4.9 _+0.4 D 8.8 _+0.5 D 65.4 ñ 1.2 C 20.4 -+ 1.0 C 97.8 -+ 8.3 A 215.3 _+9.6 C 92.5 _+12.5 A (151) (155) (151) (149) (149) (154) (98) --

Data from juveniles only. Femalenot radio-taggedand excludedfrom analyses. distancebetween daily roostsites typically in- 1-5 after fledging, adults (the male, the female, creasedduring weeks 6-8 (Fig. 5c), with sig- or both) roostedwith young 67.5%of the time nificant increasesduring this period in four (Fig. 6). In contrast,adults roostedwith young families. only 30.6%of the time during weeks 6-9 post- The mean distance between roost sites of fledging. Siblingswere also lesslikely to roost adults (both male and female) and young did with eachother during the weeksimmediately not vary significantlyduring weeks 1-5 (Fig. 5: before dispersal. In four families with three d, e). However, distancesincreased significant- young,all siblingsroosted together 86.5% of the ly during either week 6 or 7 in all families,and time during weeks 1-5 after fledging. During remainedsignificantly higher until young dis- weeks 6-9, all siblings roosted together only persed from natal areas. 47.5% of the time. All three siblings roosted Associationof family members.--Duringweeks separatelyonly 1.5%of the time during weeks

TABLœ3. Comparisonsof roostsites (LS œ+ SE) usedby adult and juvenile EasternScreech-Owls (n = 6 families)and by adult malesand females(n = 4 families)during the postfledgingperiod. Sample sizes are in parentheses(*** = P < 0.001; NS = P > 0.05).

Adults Juveniles Adult male Adult female Roost type Open limb (%) 47.0 (205) 46.0 (308) NS 54.1 (85) 48.1 (76) NS Tangle (%) 34.9 (152) 37.0 (248) 33.8 (53) 33.5 (53) Conifer (%) 18.1 (79) 17.0 (114) 12.1 (19) 18.4 (29) Roost height (m) 10.6 + 0.3 10.0 + 0.3 NS 12.2 + 0.5 12.7 + 0.5 NS (417) (651) (149) (152) Tree height (m) 14.7 + 0.4 13.8 + 0.3 NS 16.5 + 0.6 17.0 + 0.6 NS (433) (670) (157) (156) Percent height (%) 74.0 + 0.8 72.4 + 0.6 NS 75.4 + 1.1 75.0 + 1.1 NS (415) (650) (148) (152) DBH (cm) 24.8 + 0.6 23.1 + 0.5 NS 26.2 + 1.0 26.8 + 1.0 NS (432) (662) (157) (156) Dist. from nest (m) 257.8 + 5.8 249 + 4.6 NS 206.3 + 13.2 211.8 + 13.3 NS (436) (672) (159) (157) Dist. from previous 78.5 + 5.8 48.7 + 4.5 *** 53.5 + 6.4 50.8 + 6.6 NS roost (m) (327) (534) (112) (108) 574 BELTHOFFAND RITCHISON [Auk, Vol. 107

TABLE4. Roostcharacteristics (LS œ_+ SE) of accompanied(by at leastone adult) and unaccompaniedjuveniles in five EasternScreech-Owl families (we excluded the GoosePen family becausethe adult female could not be locatedconsistently). Sample sizesare in parentheses;* P < 0.05; ** P < 0.01; *** P < 0.001; NS P > 0.05.

Family Off-property Muddy Creek Variable Accomp. Unaccomp. Accomp. Unaccomp. Roost height (m) 9.6 + 0.5 9.7 + 0.6 NS 11.9 + 0.5 8.1 + 0.6*** (55) (49) (61) (46) Tree height (m) 13.2 + 0.7 13.6 + 0.8 NS 15.5 + 0.7 11.7 + 0.8*** (55) (49) (61) (46) DBH (cm) 26.1 + 1.7 27.1 + 1.8 NS 27.6 + 1.6 21.2 + 1.9' (55) (49) (61) (46) Percent height (%) 73.3 + 2.0 71.3 + 2.1 NS 76.5 + 1.9 69.2 + 2.2* (55) (49) (61) (46) Dist. from roost (m) 78.7 + 10.6 44.3 + 15.6 NS 35.9 + 11.8 28.1 + 13.1 NS (41) (19) (33) (27) Dist. from nest (m) 407 + 14.8 402 + 15.7 NS 107 + 14.1 103 + 16.2 NS (55) (49) (61) (46)

1-5 after fledging, and 7.5% of the time during young owls dispersed(mid-July). As in other weeks 6-9. Thus, when all three siblings did areas(VanCamp and Henny 1975, Merson et al. not roost together, usually only one sibling 1983, Smith et al. 1987), Eastern Screech-Owls roostedalone. One sibling roostedapart from in Kentucky frequently roost in tree cavities the other two 11.7%of the time during weeks during the autumn and winter. The increased 1-5 postfledging,and 45% of the time during use of cavities during the autumn and winter weeks 6-9. One family (Stream) had only two in part reflectschanges in the amount of cover young,which roostedtogether 96.4% of the time (VanCamp and Henny 1975). Cavities may also during weeks 1-5 after fledging,and 78.9%of provide a more favorable microclimate (Mc- the time thereafter. Comb and Noble 1981, Smith et al. 1987) and, therefore, even those trees that provide cover DISCUSSION throughout the year (e.g. eastern red cedar) are rarely usedduring late autumn and winter (pers. EasternScreech-Owls typically selectedroost obs.). sites that provided concealment. Trees fre- Screech-Owlsroosted in treesthat were sig- quently usedfor roostingwere often thosewith nificantly shorter than the average available. densefoliage. In contrast,available trees rarely This apparent preference for shorter trees prob- or never used for roosting appearedto provide ably results in part from the frequent use of little cover.Tangles (vines) alsoprovided cover easternred cedarsand tangle roosts.Eastern red and were frequently used for roosting. In ad- cedarssampled on the area were significantly dition to concealingbirds from potential pred- smaller than deciduous trees. Also, some of the ators, the dense cover of most roost sites prob- vines that createtangle roosts(e.g. Loniceraja- ably provided favorable microclimates(e.g. ponica)do not grow very high (Wharton and shadeand shelter from precipitation). Hayward Barbour 1973), and are more common in shorter and Garton (1984) found that Western Screech- trees. Screech-owls may also use lower roost Owls (Otuskennicottii), Boreal Owls (Aegoliusfu- sitesduring summer becausesuch sites are cool- nereus),and Northern Saw-whet Owls (A. acad- er and provide more shelter from wind than icus) prefer roosts that provide both thermal higher sites in the canopy. protection and concealment. We found that individuals usually did not use We rarely found screech-owlsin tree cavities. the sameroost site on successivedays. Our pres- However, our observations ended when the ence may have disturbed the owls and caused July 1990] Eastern Screech-Owl Roosts 575

TABLE 4. Extended.

Family Trap Range Stream Hilltop Accorap. Unaccomp. Accorap. Unaccomp. Accorap. Unaccomp. 21.6 + 0.4 18.6 + 1.1' 15.0 + 0.6 4.6 + 0.6*** 5.2 + 0.7 4.8 + 0.5 NS (82) (13) (53) (46) (31) (66) 25.3 ___0.6 22.5 + 1.5' 20.2 ___0.7 9.1 ___0.8*** 7.6 + 1.0 9.1 + 0.7 NS (82) (13) (59) (47) (31) (67) 35.0 + 1.4 27.5 + 3.5 NS 28.7 + 1.7 14.7 + 1.9'** 16.1 + 2.3 19.3 + 1.6 NS (82) (13) (59) (45) (30) (63) 85.1 + 1.6 77.6 + 4.1 NS 77.8 + 2.0 59.8 + 2.2*** 71.2 + 2.7 62.5 + 1.8'** (82) (13) (53) (45) (31) (66) 16.3 + 9.3 17.1 + 34.0 NS 38.2 + 9.0 64.0 + 9.8 NS 47.4 + 12.2 62.5 + 8.5 NS (53) (4) (57) (48) (31) (64) 276 + 12.1 311 + 30.5NS 85 + 14.3 111 + 15.9NS 223 + 19.7 180 + 13.4NS (82) (13) (59) (48) (31) (67)

them to seek different roostseach day. Alter- trees with limited understory,and owls in this natively, using different roost sites each day territory roostedsignificantly higher in signif- mayreduce the chancesof predation,especially icantly larger (taller and with greater diame- on the more vulnerableyoung. We often noted ters) trees than birds in other territories. Habitat "whitewash" beneath roost sites, even those differences also contributed to differences in used for just one day. Reuse of sites, and the the mean distance between the nest and roost accumulationof droppings,could attract poten- sites.Two families(Off-property and GoosePen) tial predators.It is also possiblethat limited used nest cavities located in relatively small, reusesimply indicatesthat many suitableroost isolated woodlots adjacent to areas character- sitesare available. Both Merson et al. (1983) and ized by woodedfence rows and scatteredpatches Smith et al. (1987) reported repeateduse of cer- of treesseparated by old fields and cropland. tain roost sitesby screech-owls,and similar be- Becausethese families rarely roostedin the small havior has been reported in other owl species woodlots that contained their nests and sur- (e.g.Dunstan 1970, Barrows 1981, Hayward and rounding areaswere sopatchy, the averagedis- Garton 1984). However, repeated use of roost tance from nest to roost sites increased. In these sitesby EasternScreech-Owls is more common habitats, suitable roost sites were also more scat- during winter than in summer, and often in- tered, which contributed to the greater dis- volves the repeated use of certain tree cavities tances between successive roosts in some fam- or nest boxes (Merson et al. 1983, Smith et al. ilies. 1987).Reduced cover from leaf fall during the We found no differencesin the roostingbe- autumn months plus the use of favorable mi- havior of adult male and female screech-owls, croclimatesduring cold winter months proba- and differencesbetween adults and juveniles bly limits the number of suitable roost sites occurred in one or two families. Because fami- available to screech-owlsduring winter. This lies often roosted together, few differences couldlead to the repeateduse of especiallyfa- amongfamily membersmay be expected.Even vorable sites (e.g. certain cavities). when young owls roosted away from parents, Although we noted significantvariation in however, we observed no consistent differ- roosting behavior among families, this varia- ences.It appearsthat, regardlessof either sex tion appeared to result from differences in the or age (with the possibleexception of newly areasthese families occupied. For example,one fledged owls as described below), Eastern territory (Trap Range) contained tall stands of Screech-Owlsexhibit similarroosting behavior. 576 BELTHOFFAND RITCHISON [Auk, Vol. 107

[] MuddyCreek Male [] MuddyCreek Female 1006080 t[]Off-property MaleFemale 40- 20-

0 ß Juvenile 1 Juvenile 2 Juvenile 3 Juvenile I Juvenile 2 Juvenile 3

1oo [] TrapRange Male [] TrapRange Female 80- [] StreamFemale

60-

20 20

0. Juvenile 1 Juvenile 2 Juvenile 3 Juvenile 1 Juvenile 2 Fig.3. Coefficientsof associationin roost sites between adult Eastern Screech-Owls and each of their offspring in four families.

As the postfledgingperiod progressed,owls significantvariation during the postfledging usedtangle roost sites more and openlimbs less. period, but we observedfew consistenttrends. At leasttwo factorscontributed to this apparent Screech-owlsroosted significantly closer to nest shift in preferences.Most important,tangle sites sites during the first week after fledging. For probablybecome more commonas the summer progressesand vinescontinue to grow,and tangles probablyincrease in quality (amountof cover). A secondfactor may be the increasing age of young owls. During the first few days after leaving nestcavities, young screech-owls may use roostsites unlike thosetypically used by adults(Gehlbach 1986). Sometimes they roost on open limbs that provide little or no cover (pers.obs.). Young screech-owls may usesuch sitesbecause they are unable to fly and are lim- 20' ited in their choice of roost sites, or because somelearning is involvedin siteselection. Open limbs may also provide adult owls coming to i 2 3 4 5 6 7 8 feed young with easieraccess, and the young owlsmay be unableto moveto bettersites after Week Postfiedging being fed. Fig.4. Screech-owlroost type use during the post- Other roost-site characteristics also exhibited fledging period (familiespooled). July 1990] EasternScreech-Owl Roosts 577

...... TrapRange ...... S• / ...... GoosePen ..... I-Im•p / 20'

2 3 4 5 6 7

1000.

900'

800- ...... Slre•m ...... Goose Pen 700' ...... Goose Pen

600-

500.

400. /...;...... ,,..': 300-

200'

100'

0 :-- , d. 2 3 4 5 6 7

1 2 3 4 5 6 7 • Week Posfiledging Fig. 5. Weeklyfamily means (adults and juveniles pooled) for (a) roostheight, (b) distancefrom nest, (c) distancefrom previous roost, (d) distancebetween young and adult male, and (e) distancebetween young and adult female. severaldays after leavingthe nestcavity, many for adultsand juvenilesin four families.During young screech-owlsare unableto fly (Gehlbach this same week, the distance between roost sites 1986).Thus, young owls, and the adults roost- of adults and young increasedsignificantly. ing with them, remained closer to nest sites Subsequently, adults and their young rarely during the first week after fledging. roosted together. The increased distancesap- In week 6 after leaving the nest, distances pear to coincide with increasedindependence betweendaily roostsites increased significantly of the young. Basedon biweekly home-range 578 BELTHOFFAND RITCHISON [Auk, Vol. 107

10o after fledging supportsthis. In contrast,Link- hart and Reynolds(1987) found that broodsof Flammulated Owls divided into two subgroups, one attended only by the adult male and the other only by the adult female. Flammulated Owl subgroupsmoved from nestsin different directions, and subsequent contact between them was not observed during the remainder of the fledgling dependencyperiod (Linkhart and Reynolds 1987). Despite the difficulty of observingyoung birds after they leave the nest,

o brood division has been reported in a wide va- 1 2 3 4 5 6 7 riety of species(McLaughlin and Montgomerie Week Postfledging 1985). In fact, McLaughlin and Montgomerie (1985) could find no clear examplesof species Fig. 6. Percentageof juvenileroost sites at which one or more adult owls were present (i.e. accom- in which parentsdid not divide their broods. panied young). Our results and those of Gehlbach (1986) suggest that screech-owlbroods are not divided into subgroups,but more detailed observations overlap between adult owls and their young, are needed. For example, McLaughlin and juvenile owls begin to move more widely out- Montgomerie(1985) observed that both parents sidethe rangesof their parentsat this time (Belt- fed young Lapland Longspurs(Calcarius lap- hoff 1987). Linkhart and Reynolds (1987) re- ponicus)on the day of fledging,and that brood ported similar observationsin a family of division occurredon the secondday after fledg- FlammulatedOwls (Otusfiammeolus). The mean ing. Thus,both adultsmay feed young screech- distance between roosts of an adult male and owls soon after fledging (i.e. Gehlbach 1986), its young was lessthan 35 m during the first 13 and brood division may occur later in the post- days after fledging, but it increasedto 227 m fledgingperiod. Further, Boxall (1983) observed during the next 20 days(Linkhart and Reynolds that eachadult fed particularyoung in a family 1987). Forsmartet al. (1984) also reported that of American Redstarts(Setophaga ruticilla), even adult Spotted Owls (Strix occidentalis)roosted though the brood remained together. There- with their young lessfrequently as the young fore, further observations of adults feeding matured. younglater in the postfledgingperiod are need- Although we found that the frequencywith ed to determine if brood division occurs in East- which juvenile screech-owlsroosted with sib- ern Screech-Owls. lings declinedsomewhat during the postfledg-

ing period,siblings frequently roosted together ACKNOWLEDGMENTS up to the time of dispersalfrom natal areas. Spotted Owl siblings also remain together We thankPaul Cavanagh, Keith Krantz, Earl Sparks, throughout the summer (Forsmartet al. 1984). and Tim Towles for assistance in the field, and Dale Although suchbehavior suggeststhat screech- Droge,David Tonkyn, and SteveWagner for sugges- owl siblingsmay dispersetogether, we found tionson the manuscript.Dwight Smith and Alan H. that they do not (Belthoffand Ritchison1989). Brushoffered many constructivecomments on the submitted version. We especially thank William Advantages to siblings that continue to roost BridgesJr. for assistancewith statisticalanalyses. Fi- together after independenceare unclear, but nancialsupport was providedby SigmaXi, the Sci- they may relate to those associatedwith living entific ResearchSociety, and by EasternKentucky in groups. University.Clemson University provided computer Male and female Eastern Screech-Owls roost- accessfor many of the statisticalanalyses. ed with each of their young a similar proportion of time. This behavior and the tendency of sib- LITERATURE CITED lings to roosttogether suggest the absenceof brood division in screech-owls. Gehlbach's BARROWS,C. 1981. Roostselection by SpottedOwls: (1986) observation of a fledgling screech-owl an adaptationto heat stress.Condor 83: 302-309. being fed by both adults on the secondnight BELTHOFF,J. R. 1987. Post-fledgingbehavior of the July1990] EasternScreech-Owl Roosts 579

EasternScreech-Owl (Otusasia). M.S. thesis, Rich- MCLAUGHLIN, R. L., & R. D. MONTGOMERIE. 1985. mond, Eastern Kentucky Univ. Brooddivision by Lapland Longspurs.Auk 102: ß & G. RrrCHISON. 1989. Natal dispersal of 687-695. Eastern Screech-Owls. Condor 91: 254-265. MERSON, M. H., L. D. LETA, & R. E. BYERS. 1983. Ob- BOX^LL,P.C. 1983. Observationssuggesting paren- servationson roosting sites of Screech-Owls.J. tal division of labor by American Redstarts.Wil- Field Ornithol. 54: 419-421. son Bull. 95: 673-674. OWEN, D. F. 1963a. Variation in North American CAVANAGH, P. g., • G. RITCHISON. 1987. Variation ScreechOwls and the subspeciesconcept. Syst. in the bounceand whinny songsof the Eastern Zool. 12: 8-145. Screech-Owl. Wilson Bull. 99: 620-627. 1963b. Polymorphism in the ScreechOwl COLE,L. C. 1949. The measurementof interspecific in eastern North America. Wilson Bull. 75: 183- association.Ecology 30:411-424. 190. DUNSTAN,T. C. 1970. Post-fledgingactivities of ju- RITCHISON,G., P.M. CAVANAGH,J. R. BELTHOFF,&; E. venile Great Horned Owls as determinedby ra- J.SPARKS. 1988. The singing behaviorof Eastern dio-telemetry. Ph.D. dissertationß Vermillion, Screech-Owls:seasonal timing and responseto Univ. South Dakota. playback of conspecificsong. Condor 90: 648- FORSMAN, E. D., E. C. MESLOW, & H. M. WIGHT. 1984. 652. Distributionand biology of the SpottedOwl in Ross,A. 1969. Ecologicalaspects of the food habits Oregon. Wildl. Monogr. 87: 1-64. of insectivorous Screech Owls. Proc. West. Found. GEHLB^CH,F.C. 1986. Odd couplesof suburbia.Nat. Vert. Zool. I: 301-344. Hist. 6: 56-66. SMITH, D. G., & R. GILBERT.1981. Backpack radio GREIG-SMITH,P. 1964. Quantitative plant ecology. transmitter attachment success in Screech Owls London, Butterworth. (Otus asia). N. Am. Bird Bander 6: 142-143. HA'SWARD, G. D., & E. O. GARTON. 1984. Roost habitat , A. DEVINE, & R. GILBERT. 1987. Screech Owl selectionby three small forestowls. Wilson Bull. roost site selection. Birding 19: 6-15. 96: 690-692. ß & R. GILBERT. 1984. Eastern Screech-Owl HEGDAL, P. L., & B. A. COLVIN. 1988. Potential hazard home range and use of suburban habitats in to EasternScreech-Owls and other raptorsof bro- southern Connecticut. J. Field Ornithol. 55: 322- difacoum bait used for vole control in orchards. 329. Environ. ToxicologyChem. 7: 245-260. TURNER,L. J., & R. W. DIMMICK. 1981. Seasonalprey HRUBANT,H.E. 1955. An analysisof the colorphases captureby the ScreechOwl in Tennessee.J. Ten- of the EasternScreech-Owlß Otus asia, by the gene nessee Acad. Sci. 56: 56-59. frequency method. Am. Nat. 89: 223-230. VANCAMP,L. F., &; C. J. HENNY. 1975. The Screech HURLBERT,S. H. 1984. Pseudoreplicationand the de- Owl: its life history and population ecology in sign of ecologicalfield experiments.Ecol. Mono- northern Ohio. North Am. Fauna No. 171. Wash- gr. 54: 187-211. ington, D.C., U.S. Fish and Wildl. Serv. KREBS,C.J. 1989. Ecologicalmethodology. New York, v^N DERWEYDEN, W.J. 1975. Scopsand screechowls: Harper and Row. vocal evidencefor a basicsubdivision in the ge- LINKHART, B. D., & R. T. REYNOLDS. 1987. Brood di- nus Otus (Strigidae). Ardea 63: 65-77. vision and postnestingbehavior of Flammulated WHARTON, M. E., & R. W. BARBOUR. 1973. Trees and Owls. Wilson Bull. 99: 240-243. shrubsof Kentucky. Lexington, Univ. Kentucky MARSH^L•, J. T., JR. 1967. Parallel variation in North Press. and Middle American Screech Owls. West. Found. ZAR, J. H. 1974. Biostatisticalanalysis. Englewood Vert. Zool., Monogr. 1: 1-72. Cliffs, New Jersey,Prentice-Hall Inc. MCCOMB, W. C., & R. E. NOBLE. 1981. Microclimates of nest boxes and natural cavities in bottomland hardwoods. J. Wildl. Manage. 45: 284-289.