j. RaptorRes. 35(3):214-220 ¸ 2001 The Raptor ResearchFoundation, Inc.

HABITAT USE, POPULATION DENSITY, AND HOME RANGE OF ELF (MICRATHENE WHITNEYI) AT SANTA ANA NATIONAL WILDLIFE REFUGE, TEXAS

CHRISTOPHER M. GAMEL 1 AND TIMOTHY BRUSH Departmentof Biology,University of Texas-PanAmerican, Edinburg, TX 78539 U.S.A.

ABSTRACT.--Wecollected data on the habitat use, home range size, and population density of the Elf (Micrathenewhitneyi idonea) in the SantaAria National Wildlife refuge (SANWR) Lower Rio Grande Valley, Texas. Fourteen nocturnal surveysin 1995 and 1996 indicated that Elf Owls used habitat (92%) more than riparian woodlands(8%). Habitatsused had high foliage densities(greatest density = 2.5-3.0 m) with partial canopies (x = 3.8 +_ 0.36 m [+_SE] and semi-open understories (greatestdensity <1 m). Unused chaparral habitat lacked high canopy coverageand had a denser understory,while riparian woodlandshad greater canopyheights (• = 5.25 m) and open understories. Home range size determined by radiotelemetry averaged1.0.5 +_ 0.33 ha (range = 0.24-2.60, N = 9). We estimatedthe maximum potential population size in SANWR to be 802 Elf Owls, assuminga home range size of 1.05 ha per breeding pair and saturationof preferred habitat. KEY WORDS: Elf Owl; Micrathene whitneyi; homerange,, habitat selection;chaparral; Santa Ana National WildlifeRefuge,, Texas.

Uso de habitat, densidadpoblacional y rango de hogar de Micrathenewhitneyi en el Refugio Nacional de Vida Silvestre de Santa Aria, Texas R•st3MF•N.--Colectamosdatos sobre el uso de habitat, tamafio del rango de hogar y densidadpoblacional de Micrathenewhitneyi iddnea en el Refugio Nacional de Vida Silvestrede Santa Ana en el valle bajo del Rio Grande en Texas. (RNVSSA) Catorce monitoreosnocturnos hechos en 1995 y 1996 indicaron que Micrathenewhitneyi utiliz6 el habitatde chaparral(92%) masquelos bosques riberefios (8%). Loshabitat utilizadostenJan densidades de follaje mayores(mayores densidades = 2.5-3.0 m, con dosdesparciales (• = 3.8 +_0.36 m _+SE) y con sotobosquessemiabiertos (con una densidadde < 1 m) . El habitat de chaparral no utilizado no tuvo una cobertura de dosel alta con una densidad mayor en el sotobosque, mientras que los bosquesriberefios tuvieron una mayor ahura del dosel (• = 5.25 m) y un sotobosque abierto. E1 tamafio del rango de hogar fue determinado por el promedio de los resultadosde telemetrfa 1.05 +_0.33 (rango = 0.24-2.60 ha, N = 9) . Estimamosun tamafiopotencial m•ximo de poblaci6nen el RNVSSA de 802 bfihos enanos,asumiendo un rango de hogar dc 1.05 ha por pare.ja reproductivay una saturaci6n del hfibitat preferido. [Traducci6n de CfisarMfirquez]

The Elf Owl (Micrathenewhitneyi) extends fi:om 1983). M. w. idoneabreeds along the Texas-Mexico the southwestern United States to southern Mexi- border (Wauer 1971) and patchily in Tamaulipas co. The subspeciesM. w. whitn(yiis the most exten- (E Gehlbach pets. comm.). M. w. whitneyioccurs sively studied of the fbur subspeciesbreeding in in a variety of habitat types including evergreen southern Arizona (Ligon 1968, Goad and Mannan woodlands, Sinaloan deciduous forests (Ligon 1987), extreme southeastern California (Rosen- 1968), and Sonoran desert in associationwith sa- berg et al. 1991), western New Mexico, the Big guaro cactus (Cereusgiganteus, Goad and Mannan Bend region of Texas(Van Tyne and Sutton 1937), 1987). Because the habitats in southern Texas dif: and into western Mexico (Ligon 1968, Staceyet al. t(r drastically,habitat requirements of M. w. idonea are probably different. The Elf Owl is a breeding resident in the Lower 1Present address: Marine PhysiologyLaborato- Rio Grande Valley (LRGV; Cameron, Willacy, Hi- ry, 5007 Ave. U, TexasA & M University,Fort Crockett, dalgo, and Starr counties) of south Texas (Ober- Galveston, TX 77551 U.S.A. holser 1974, Gehlbach 1987). The LRGV delta •s

214 SEPTEMBER 2001 HABITAT USE OF ELF OWLS 215

the alluvial floodplain of the Rio Grande as it flows variation in the density and habitat use of the owls. The direction of searches was reversed on four occasions to toward the Gulf of Mexico. Much of the vegetation reduce potential biases. occurring in the area is adapted to the annual Along the routes, we broadcast tape-recorded calls of flooding of the river system.Since its completion Elf Owls every 300 m and listened for 1 rain for sponta- in 1953, Falcon Dam has affected the flora and neously calling Elf Owls, following which the recorded fauna in the floodplain. Today, seasonalflooding call was played lwo more times for about 15 sec (6 calls) Each presentation of calls was followed by 1 rain of lis- of the Rio Grande rarely occursand there hasbeen tening and a search of the immediate vicinity with a a gradual changefrom riparian woodlandsto chap- hand-held flashlight for owls.The estimatedlocations of arral communities, locally known as Tamaulipan all vocalizingowls were recordedon 7.5 rain topographic thornscrub (Vora 1990). The Elf Owl was first doc- mapsof the area (Camel 1997). umented in the LRGV in 1889 (Sennerr 1889). For We captured Elf Owls in mist nets. Two to four, 3 X 13 m mist nets were set up shortly after sunset in areas the next 70 yr, it wasbelieved to be extirpatedfrom where Elf Owls had been heard calling. Continuous the LRGV, until it was rediscovered in 1960 by broadcastingof a recorded Elf Owl call was used to lure James and Hayse (1963) who found it in the west- owls into the mist nets. Owls that did not appear to be ern portion of the LRGV in Starr and western Hi- overly stressedwere fitted with 1.2-1.4 g radio transnut- ters (Wildlife Materials, Inc., Carbondale, IL U.S.A.). In dalgo counties. Oberholser (1974) found two 1995, four transmitters were attached with a backpack fledglings in Santa Ana National Wildlife Refuge style cross-chestharness. In 1996, five transmitters were (SANWR) in 1965 and Gehlbach (1987) found directly attached to the dorsal feathers along the spine, them nesting at the edge of evergreen forests in as per Warnock and Warnock (1993). SANWR fi:om 1973-78 suggestingthat the species Owl locationswere determined by triangulation with a hand-held receiver (TRX-10S, Wildlife Materials, Inc, wasspreading east. Carbondale, IL U.S.A.) and a three element ¾agi anten- Since 1920 there has been an overall shift in the na. Bearingswere taken by a single individual, with read- habitat composition at SANWR so that it now con- ings taken from pre-establishedstations. Tracking usually sistsmainly of chaparral (Vora 1990). Today,ripar- began within one hour of sunset and ranged from 1-5 ian woodlandscover only about 25% of SANWR hr in length. Tracking was repeated severaltimes over a month period until either 30 locationswere recorded for (E. Hopson pets. comm.). The change in habitat each owl, the transmitter battery failed, or the owl left has been directly attributed to the absenceof sea- the area. Area/observation curves indicated 30 locations sonal flooding of the Rio Grande (Ramirez 1986, were adequate to show movement patterns of the Elf Vora 1990), but few studies have looked at subse- Owls over a month-long portion of the breeding season Individuals were located at 30 min intervals and all bear- quent changesin the plant and communities. ings for a single location were collectedwithin a 10-nun The objectivesof this study were to describe pat- period to reduce error. Home range sizeswere calculated terns of habitat use, home range size, and popu- with the TELEM88 program (Coleman 1989) using a lation densityof the M. w. idoneasubspecies of the 100% minimum convexpolygon. No individualswere fol- Elf Owl in the wildlife refuge. lowed in both 1995 and 1996, as no owlswere recaptured We identified three habitat categories:chaparral habitat STUDY AREA AND METHODS utilized by Elf Owls (CWO), chaparral habitat not utilized by Elf Owls (CWNO), and riparian woodlands.An analysm The study area consistedof 842 ha of native chaparral of the vegetation in each habitat was conducted using a and other wooded habitat in southern Hidalgo County, point-quartersampling method (Brower et al. 1977).Sam- TX. The refuge encompassesthe largestremaining tracts ple pointswere placed at 10-m intervalsalong a 100-mtran- of chaparral and riparian woodland in the LRGV (Vora sect. Within each habitat, three transects were established 1990). A tour road, 10 km in length, and an extensive yielding a total of 33 samplepoints. In the CWNO and •n trail systemcut through both vegetativetypes and provid- the riparian woodlands,starting locations and directionsfor ed a route for surveyingthe area for Elf Owls. all transectswere chosenrandomly. Within the CWO, tran- We conducted nocturnal surveysduring the breeding sectswere establishedwithin the home range of three ran- seasons(late March-early August) of 1995 and 1996 to domly picked, radio-trackedowls. At each sample point, determine relative owl densities and second-order habitat measurementswere taken on the nearestwoody species use. The surveyroute had 7.26 km of chaparral (67%) m in height in each quadrant.For each individualwoody and 3.74 km of riparian woodland (33%) in the 1995 species,point-to-plant distance, diameter at breast height field season.The surveyroute was increased in 1996 in- (DBH), and individualtree canopycoverage (at the widest creasingthe amount of chaparral and riparian woodland point of coverage)were recorded.Canopy height wasmea- to 7.32 km (60%) and 4.88 km (40%), respectively.Sur- suredand % canopycover was estimated with a densiome- veyscovered 68% of SANWR. ter. Foliagedensity was quantified using a modifiedversion Surveyswere made on foot while listening for owl re- of Mills et al. (1991) by marking a 3-m extendingpole at sponsesto recorded playbacks.They were conducted 0.5 m intervalsalong the entire length.When fullyextended once every three weeksto determine if there wasseasonal and held at arm's length abovethe head, total height cow 216 CAMEL AND BRUSH VOL. 35, NO. 3

Table 1. Comparisonof the meanstructural composition of the vegetationamong habitats in the SantaAna National Wfidlife Refuge,Texas. Characteristics included are tree density,individual tree cover,diameter at breastheight (DBH), % canopycover, and canopyheight. An asterisk(*) betweentwo habitatsindicates a significantdifferent (P < 0.05).

RIPARIAN CHAPARRAL WITH CHAPARRAL WITH CHARACTERISTIC WOODLAND OWLS NO OWLS Tree density(m '•) 0.32 + 0.10 0.41 + 0.07 0.53 + 0.09 Tree cover (m) 4.07 + 0.27 * 3.35 + 0.22 * 2.61 + 0.15 DBH (cm) 12.12 + 1.16 13.04 + 1.55 11.96 + 1.59 % Canopycover 67.70 + 0.06 55.20 + 0.06 40.00 + 0.06 Canopyheight (m) 4.96 + 0.28 * 3.81 + 0.36 2.82 + 0.30 ered was5.5 m. At eachsample point, the polewas erected 79% similar, and riparian woodlandsand CWNO verticallyand the numberof plant toucheswere recorded 64% similar. Sorensen'scoefficients comparing the per 0.5 m segment.Records included touches by all vege- transects in the CWO to each other indicated with- tative matter. We used SPSS-X (SPSS,Inc., Chicago, IL U.S.A.) for in-habitatdissimilarity to be as great as 26%. statisticalanalyses. Results were considered significant Riparian woodlands had significantly greater when P --<0.05. Means are accompaniedby standarder- canopyheights (F = 10.219, df = 2, P < 0.001; rors on all measurements.A chi-squaretest wasused to determine differenccs in habitat use between the 1995 Table 1) and individual tree covers (F = 11.425, df and 1996 field seasons.For each survey,the expected = 2, P < 0.001) than CWO. Differences in tree d•stribution of Elf Owls reflected the area covered in ri- density,DBH, and % canopycover were not sig- parian woodlandsand chaparral. A Sorensen'scoeffi- nificant (P > 0.05). CWO had significantlygreater cient of community similaritywas calculated to compare individual tree cover (F = 11.425, df = 2, P < the similarityof speciesbetween different habitat types. A one-wayanalysis of variance (ANOVA) wasused to ver- 0.001) than CWNO. Tree density,DBH, % canopy ify any observeddifferences between the three habitats cover,and canopyheight were not significantlydif- for the variablesquantified in the vegetationtransects. In ferent (P > 0.05). The vegetationin CWO wassig- caseswhere a significantdifference wasfound, a Tukey nificantlydenser at 2.50-3.00 m (F = 3.281, df = HSD testwas used to identify which habitatsdiffered sig- mficantly. In all cases,CWO was compared to both 2, P = 0.042; Fig. 1) than in riparian woodlands CWNO and riparian woodlands. but, in riparian woodlands,the vegetationwas sig- nificantlytaller 5.00-5.50 m (F = 6.275, df = 2, P RESULTS = 0.003). The vegetation in CWNO was denser We recorded a total of 145 Elf Owl locations. than CWO (F = 8.955, df = 2, P < 0.001; Fig. 2) Significantlymore locationswere in chaparral(N but the vegetationin CWO wassignificantly taller = 134) than in riparian (N = 11) habitatsin 1995 >5.50 m (F = 14.145, df -- 2, P < 0.001). (X2 = 52.408, df = 1, P < 0.001) and 1996 (X2 = Ten Elf Owlswere capturedin 1995 and seven•n 51.946, df = 1, P < 0.001). The three transects 1996 over the course of nearly 52 hr of mist net within riparian woodlandsincluded 13 tree species, trapping.Of these,nine Elf Owlswere instrumented of which anacua (Ehretiaanacua), sugar hackberry with radio transmitters (4 in 1995 and 5 in 1996). ( Celtislaevigata), and cedar elm ( Ulmuscrassifolia) Home range size ranged from 0.24-2.60 ha (• = dominated. CWO included 12 speciesof trees,with 1.05 + 0.33 ha). There was no correlation between la coma (Bumeliacelastrina), Texas persimmon (Di- number of locationsper owl and home range size ospyrostexana), and spinyhackberry ( Celtispallida) (r = -0.292, P = 0.20) but an owl with 30 locations the dominant species.CWNO included 16 tree spe- had the smallesthome range (0.24 ha). We did not cies,with the dominant speciesbeing spinyhack- find evidenceof home range overlap. berry, la coma, and guayacan(Guaiacum angustifol- DISCUSSION zum).Despite the difference in speciesdominance, there wasoverlap in speciespresence between the Elf Owls made significantlygreater use of chap- different habitat types.A Sorensen'scoefficient of arral habitat than riparian woodlandsat SANWR. community similarity showed riparian woodlands A total of 21 speciesof treesand shrubswere found and CWO to be 67% similar, CWO and CWNO along the nine vegetationtransects and there was SEPTEMBER2001 HABITATUSE OF ELF OWLS 217

0.00-0.50

0.50-1.00

1.00-1.50

1.50-2.00

2.00-2.50

* 2.50-3.00

3.00-3.50

3.50-4.00

4.00-4.50

4.50-5.00 Riparian * 5.00-5.50 Woodlands * > 5.50 • CWO o 2'0 3O

Percentage of Total Canopy Volume Figure1. Comparisonof theunderstory vegetation density in riparianwoodlands and chaparral habitats used by Elf Owls.Percent of totalcanopy volume is calculatedfrom the meannumber of thepole touches by vegetation for eachof thethree transects at differentheights. An asterisk(*) nextto thepole height indicates a significant difference •n densityat thatlevel. N = 33 in riparianwoodlands and chaparral habitats used by Elf Owls. a large degree of overlapin speciesamong the 1988), Eastern Screech-Owls(Otus asio, Belthoff transects. Still, Elf Owls occurred in one habitat and Ritchison 1990), and Northern Spotted Owls more than the other. The diversityof habitatsin (Strix occidentalisoccidentalis, Bias and Gutierrez which Elf Owls are known to reside (Ligon 1968, 1992). Schaeffer and Ehlers 1979, Goad and Mannan While availabilityof cavitiesis often a limiting 1987) suggeststhat the presence,or absence,of factorfor secondarycavity nesters (Brawn and Bal- particularwoody plant speciesplays a role in hab- da 1988, Petty et al. 1994), such is likely not the itat use. case at SANWR, where primary cavity nesters are The suitabilityof individual tree speciesfor ex- common (Carter 1986, Gehlbach 1994, Brush and cavationby primarycavity nesters could also affect Cantu 1998) and cavities (rot and -ex- the distributionof secondarycavity nesters like the cavated) are abundant throughout all three vege- Elf Owl. This has been shown to be the case with tative communities(T. Brush pers. obs.). M. w. whitneyiwhich tendsto favorsaguaro cacti as An analysisof the structuralfeatures of each nest sitesdue to the abundance of Gila Woodpeck- communityshowed canopy height, individual tree er (Melanerpesuropygialis) and Gilded Flicker (Co- cover, and understory densitywere important in lapteschrysoides) excavations (Goad and Mannan the use of habitatsby Elf Owls.Canopy height dif- 1987). A similar dependenceon primary cavity feredsignificantly between riparian woodlands and nesters has been documented in Flmmnulated CWO with Elf Owlsusing the intermediatecanopy Owls (Otusflammeolus, McCallurn and Gehlbach heightsfound in CWO. The combinationof a 218 CAMEL AND BRUSH VOL. 35, NO. 3

0.00-0.50

0.50-1.00

1.00-1.50

1.50-2.00

2.00-2.50

2.50-3.00

3.00-3.50

3.50-4.00

4.00-4.50

4.50-5.00

5.00-5.50, • CWNO >5.50 cwo ß 0 10 20 30

Percentageof Total CanopyVolume Figure 2. Comparisonof the understoryvegetation density in chaparral habitat used by Elf Owls and chaparral habitat not usedby Elf Owls.Percentage of total canopyvolume is a calculatedfrom the mean numberof the pole touchesby vegetationfor each of the three transectsat different heights.An asterisk(*) next to the pole height indicatesa significantdifference in densityat that level.N = 33 in chaparralhabitat used by Elf Owlsand chaparral habitat not used by Elf Owls. higher canopy and large individual tree coverage prey items.This might explainwhy they do not use causesa distinct,but partial, canopyto form in the relatively-open riparian woodlands. Alternately, CWO which is lacking in the CWNO. Gehlbach (1987) suggestedthat competitionmight Elf Owls in SANWR use habitat that consists of occur between the Elf Owl and Eastern Screech- a distinct,but partial, canopylayer at about 4 m Owl, a resident of riparian woodlandsat SANWR. and a semi-openunderstory which is most dense Ligon (19{38) mapped Elf Owl home ranges at 2.50-3.00 m. They do not use areasthat haveno based on the locationsof calling males estimated understory or have very dense understoriesand the size to be 03 ha. Our radio-tele•netrydata sup- very high or low canopies. Flammulated Owls ports the observationthat Elf Owls occupysmall (McCallum and Gehlbach 1988) and Boreal Owls home ranges. Although our home range size of (Aegoliusfunereu& Norberg 1970), are known to 1.05 ha waslarger, six of the nine owlstracked used dive steeplywhen leaving a perch, then level off areas <0.{3 ha. Of the three that occupied larger and fly 1-2 m above the ground. Similar behavior home ranges,one moved back and forth between occurs in the Elf Owl (F. Gehlbach pers. com.), two patchesof chaparral habitat that were separat- and may thereby explain why it doesnot use chap- ed by a band of riparian woodland.While this ri- arral habitat which containsa very denseundersto- parian strip was included in the convexpolygon, ry. At the sametime, someunderstory appears nec- we observedno use of the area except as a corri- essary, possibly to provide protection from dor. Even the closestecological equivalent owl spe- predators or to maintain appropriate habitat for cies, the , maintains an average SEPTEMBER2001 HABITAT USE OF ELF OWLS 219 home range size of 14.1 ha (Howie and Ritcey assistancewas provided by N. Adame, D. Ake, G. Castillo, 1987), over lg times as large. Prey abundancehas C. Rupert, J. Rupert, V. Varela, and R. Zamora. G. Proud- fbot provided technical advice.We also thank C. Crocker- been shownto stronglyinfluence home range size BedIbrd, F. Gehlbach, S. Henry, R.W. Mannan, and P of Eastern Screech-Owls (Belthoff et al. 1993) and Stacey for suggestionson an earlier draft of the manu- Northern Spotted Owls (Carey et al. 1992). Eastern script. Funds were provided by the University of Texas- Screech-Owlsin suburban settingsmaintain small Pan American Faculty Research Grant number 119123 All captures and banding were conducted under federal home rangesof only 4-6 ha (Gehlbach 1994) com- bird-banding permit number 22544. pared to sizesranging ii:om 11.9-108 ha in other parts of North America (Johnsgard 1988, Belthoff LITEPokTURE CITED et al. 1993, Gehlbach 1994). The smallhome range BIAS,M.A. ^ND R.J. GUnERPmZ.1992. Habitat associations size we observed in Elf Owls may have been due of California spotted owls in the central Sierra Neva- to the insectivorous diet of the Elf Owls and the da. J. Wildl. Manage. 56:584-595. large numbersof insect prey in chaparral habitat. BELTHOFF,J.R. AND G. RITCHISON.1990. Nest-site selec- Several authors have indicated that Elf Owls are tion by Eastern Screech-Owlsin central Kentucky. terNtoNal (Ligon 1968,Johnsgard1988). The data Condor 92:982-990. we collectedsupported the notion that this owl ac- , E.J. SP^RKS,AND G. RITCinSON.1993. Home rang- es of adult and juvenile Eastern Screech-Owls:size, tively deiknds territories. During nocturnal sur- seasonalvariation and extent of overlap.J. RaptorRes veys,responding males tended to be spacedout 27:8-15. rather then dumped close to each other. Also, no Brawn,J.D. and R.P. Balda. 1988. Population biology of instrumented owls maintained overlappinghome cavity nesters in northern Arizona: do nest sites limit ranges. In contrast, both Ligon (1968) and Gehl- breeding densities?Condor 90:61-71. bach (pets. comm.) observedhome range overlap BROW•R,J.E., J.H. Z^R, ^ND C.N. VON ENDS.1977. Field in Elf Owls in Arizona. and laboratory methods for general ecology.William Based on our findings, we estimated the maxi- C. Brown Publishers,Dubuque, IA U.S.A. mum population of Elf Owls that can be supported BRUSH,W. AND A. Cedric. 1998ßChanges in the breeding in SANWR. In making this estimate,we assumeda community of subtropical evergreen forest in the Lower Rio Grande Valley of Texas, 1970s-1990s.Texas home range size of 1.05 ha, all of the available hab- J. Sci. 50:123-132. itat is occupied, all home rangesare occupied by •¾, A.B., S.P. HORTON, AND B.L. B•SW•. 1992. North- a breeding pair, and owls do not use habitat that ern Spotted Owls: influence of prey base and land- is not suitable ibr them. Based on our estimation scape character.Ecol. Monogr. 62:223-250. that there was 421 ha of suitable habitat in CAP,T•P,, M.D. 1986. The parasitic behavior of the SANWR, we extrapolated that a maximum of 401 Bronzed Cowbird in South Texas. Condor 88:11-25 pairs, or 802 owls, could be supported in SANWR CO•F•M•N,J.S. 1989. Program documentation to TELEM' indicating that that current population is substan- a computer systemfor analyzingradio-telemetry data. tially smallerthan the maximum that can be sup- Virginia Polytechnic Institute and State University, ported. Blacksburg,VA U.S.A. Wildliik management policiesin the LRGV are CAMEL,C.M. 1997.Habitat selection,population density, currentlyibcused on reducingthe negativeimpact and home range of the Elf Owl, Micrathenewhitneyz, at Santa Ana National Wildlife Refuge, Texas. M.S. the- that halting of annual flooding hashad on riparian sis,Univ. Texas-PanAmerican, Edinburg, TX U.S.A woodlands.Recent programsat SANWR involving GEHLBACH,F.R. 1987. Natural historysketches, densities, artificial flooding have the potential of halting, and biomassof breeding in evergreen forestsof even reversing,the habitat transitionfi:om riparian the Rio Grande, Texas and Rio Corona, Tamaulipas, woodlandsto chaparral that has been underway Mexico. •lkxasJ.Sci. 39:241-251. since the completion of Falcon Dam in 1953. Some ß1994. The EasternScreech-Owl: life history,ecol- species have apparently benefitted ii:om the in- ogy, and behavior in the suburbs and countryside creasedavailability of chaparral communities.The TexasA&M Univ. Press,College Station, TX U.S.A Elf Owl, in particular,has gone ii:om a stateof pos- Go•d), M.S. AND R.W. M^NNAN. 1987. Nest site select,on sible near-extirpation to maintaining a substantial byElf Owlsin SaguaroNatfonal Monument, Arizona. Condor 89:659-662. population since 1963. HOWIE, R. AND R. RiTCEY.1987. Distribution, habitat se- ACKNOWLEDGMENTS lection and densities of Flammulated Owls in Briush We gratefullyacknowledge E. Hopson, F. Judd, and R. Columbia. Pages249-254 in R.W Nero, R.J.Clark, R.J. Lonard for their guidance throughout this study.Field I/mapton, and R.H. Hamre [EDS.], Biology and con- 220 G)dX4}•L•XNr) BRUSH VOL. 35, No. 3

servation of northern forest owls: symposium pro- Spec. Rep. SantoAria and Rio Grande Refuges,Ala- ceedings.USDA For. Set. Gem Tech. Rep. RM-142. mo, TX U.S.A. J^MES,P. AND A. H^YSE.1963. Elf Owl rediscovered in ROSENBERG,K.V., R.D. OHMART, W.C. HUNTER, AND B.W. Lower Rio Grande DeRa of Texas. Wilson Bull. 75:179- ANDERSON. 1991. Birds of the lower Colorado River 182. valley. Univ. Arizona Press, Tucson, AZ U.S.A. JOHNSGXRD,P.A. 1988. North American Owls. Smithsoni- SCI-•a•FFER,P. ^N•) S. EHLERS [EDs.]. 1979. Owls of the an Institution Press,Washington, DC, U.S.A. west:their ecologyand conservation.National Audu- L•GON,J.D. 1968. The biology of the Elf Owl, Micrathene bon Society and Western Education Center, Tiburon, whitneyi.Museum of Zoology, Univ. Michigan, No. CA U.S.A 136. SENNETT,G.B. 1889. Micropallaswhitneyi, Elf Owl, taken MCCALLUM,D.A. ANDER. GEHLBACH.1988. Nest-sitepref- in Texas. Auk 6:276. erences of Flammulated Owls in western New Mexico. S•^•E¾, P.B., D. ROSe,T^, T. EDw•s, AND N. Jos•E. Condor 90:653-661. 1983. Northeastern extension of the breeding range M•LLS,G.S., J.B. DUNN•NO,JR., ^NDJ.M. BATES.1991. The of the Elf Owl in New Mexico. Southwest. Nat. 28:99- relationship between breeding bird density and veg- 100. etation volume. Wilson Bull. 103:468-479. VANT•E, J. ANDG.M. SUTTOn.1937. The birds of Brew- NO•ERO, R.A. 1970. Hunting techniquesof Tengmalm's ster county,Texas. Misc. Publ. Mus. Zool. Univ. Mich, Owls Aegolius)'itnereus. Ornis Scand. 1:51-64. No. 37. OBERnOLSER,H.C. 1974. The bird life of Texas. Univ. Vov•, R.S. 1990. Plant communities of the Santa Aria Na- Texas Press, Austin, TX U.S.A. tional Wildlife Refuge, Texas. TexasJ.Sci. 42:115-128. PETTY,S.J., G.S. GEO•?,AND D.I.K. ANDERSON.1994. Value WXP,NOCK, N. AND S. WARNOCK. 1993. Attachment of ra- of nest boxesfor population studiesand conservation dio-transmittersto sandpipers:review and methods. of owls in coniferous forests in Britain. J. RaptorRes. WaderStudy Group Bull. 70:28-30. 28:134-142. W^UF•R,R.H. 1971. Ecologicaldistribution of birds of the 1L•vn•z, P. 1986. Water development projects in the Rio Chisos Mountains, Texas. Southwest.Nat. 16:1-29. Grande and their relationship to the Santa Aria and Rio Grande Valley National Wildlife Refuges, Texas. Received 23 .June 2000; accepted 20 April 2001