during Autumn Migration in the Western United States...

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofitpublishers, academic institutions, research libraries, and research funders in the common goal of maximizing access tocritical research.

Morphometric Variation in Flammulated Owls Capturedduring Autumn Migration in the Western United StatesAuthor(s) :Jeff P. Smith, John P. De Long, Lynda L. Leppert, Sarah L. Stock,Gregory S. Kaltenecker, and Jay D. CarlisleSource: Journal of Raptor Research, 46(1):109-120. 2012.Published By: The Raptor Research FoundationDOI: http://dx.doi.org/10.3356/JRR-10-106.1URL: http://www.bioone.org/doi/full/10.3356/JRR-10-106.1

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in thebiological, ecological, and environmental sciences. BioOne provides a sustainableonline platform for over 170 journals and books published by nonprofit societies,associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated contentindicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should bedirected to the individual publisher as copyright holder.

http://dx.doi.org/10.3356/JRR-10-106.1

http://www.bioone.org/doi/full/10.3356/JRR-10-106.1

http://www.bioone.org

http://www.bioone.org/page/terms_of_use

http://www.bioone.org/page/terms_of_use

MORPHOMETRIC VARIATION IN FLAMMULATED OWLS CAPTUREDDURING AUTUMN MIGRATION IN THE WESTERN UNITED STATES

JEFF P. SMITH1 AND JOHN P. DELONG2

HawkWatch International, 2240 South 900 East, Salt Lake City, UT 84106 U.S.A.

LYNDA L. LEPPERT3

Raptor Research Center, Boise State University, 1910 University Drive, Boise, ID 83725 U.S.A.

SARAH L. STOCK,4 GREGORY S. KALTENECKER AND JAY D. CARLISLEIdaho Bird Observatory, Department of Biological Sciences, Boise State University, 1910 University Drive, Boise,

ID 83725 U.S.A.

ABSTRACT.—Knowledge of the migration geography of Flammulated Owls (Otus flammeolus) is limited. Wecombined data from multiple studies in the western United States to evaluate patterns of variation in Flam-mulated Owl morphometrics. Measurements from autumn migration study sites in New Mexico, Nevada, andIdaho followed a geographic cline from southeast to northwest across the species’ range. Our objectives wereto: (1) describe age- and sex-specific variation in the morphometrics of Flammulated Owls captured at thethree sites, and (2) discuss the implications of this variation for understanding the species’ migration geog-raphy. Based on a hierarchical series of factorial two-way ANOVAs, we discovered significant overall sex-specificdifferences in wing chord, exposed culmen length, and hallux length, and a consistent pattern of site-specificdifferences for most measurements. Most measurements increased from southeast to northwest, consistentwith previous data on clinal variation in the species. Evidence of significant variation in the morphometrics ofowls captured at the three migration sites suggests little longitudinal mixing of migrants in the region.

KEY WORDS: Flammulated Owl; Otus flammeolus; Bergmann’s rule; DNA-based sexing; migration geography;morphometrics; sexing; sexual size dimorphism; western United States.

VARIACION MORFOMETRICA EN INDIVIDUOS DE OTUS FLAMMEOLUS CAPTURADOS DURANTELA MIGRACION DE OTONO EN EL OESTE DE ESTADOS UNIDOS

RESUMEN.—El conocimiento sobre la geografıa de la migracion de Otus flammeolus es escaso. Combinamosdatos de varios estudios del oeste de Estados Unidos para evaluar los patrones de variacion en la morfome-trıa de O. flammeolus. Las medidas registradas en los sitios de estudio de la migracion otonal en NuevoMexico, Nevada, e Idaho presentaron un gradiente geografico desde el sudeste hacia el noroeste del rangode la especie. Nuestros objetivos fueron: (1) describir la variacion especıfica ligada a la edad y al sexo en lamorfometrıa de individuos de O. flammeolus capturados en los tres sitios, y (2) analizar las implicancias deesta variacion para entender la geografıa migratoria de esta especie. Basados en una serie jerarquica deANOVAs factoriales de dos vıas, descubrimos diferencias significativas generales vinculadas al sexo en lacuerda del ala, largo expuesto del culmen y largo del halux, y un patron consistente de diferenciasvinculadas al sitio para la mayorıa de las medidas. La mayorıa de las medidas aumentaron desde el sudestehacia el noroeste, de modo consistente con datos anteriores del gradiente de variacion en esta especie. Laevidencia sobre la variacion significativa en la morfometrıa de los individuos de O. flammeolus capturados enlos tres sitios de migracion sugiere poca mezcla longitudinal de migrantes en la region.

[Traduccion del equipo editorial]

1 Present address: H.T. Harvey and Associates, 983 University Avenue, Building D., Los Gatos, CA 95032 U.S.A.; emailaddress: [email protected] Present address: Department of Ecology and Evolutionary Biology, Yale University, New Haven, CT 06520 U.S.A.3 Present address: Georgia Aquarium, VSCM, Atlanta, GA 30313 U.S.A.4 Present address: Division of Resources Management and Science, Yosemite National Park, 5083 Foresta Road, P.O. Box700, El Portal, CA 95318 U.S.A.

J. Raptor Res. 46(1):109–120

E 2012 The Raptor Research Foundation, Inc.

109

Little is known about the biogeography and mi-gration ecology of the Flammulated Owl (Otus flam-meolus). As a small nocturnal insectivore restrictedto montane forest habitats (McCallum 1994), theFlammulated Owl has been difficult to study, espe-cially during migration. Although the species ap-pears to be more common than once believed,Partners in Flight placed the Flammulated Owl onits Species of Continental Importance Watch Listbecause of concerns about its restricted distributionand low population size (Rich et al. 2004). Flammu-lated Owls were once thought to be nonmigratory(Phillips 1942), but the species is now commonlyconsidered an obligate, medium- to long-distance,north-south migrant (Banks 1964, Ligon 1968,Winter 1974, Balda et al. 1975, McCallum 1994).

Flammulated Owls breed in montane forests ofwestern North America and Mexico (McCallum1994). Overall, the species appears to increase insize from the southeastern to the northwesternparts of its range (Bergmann’s rule; Marshall 1967;Blackburn et al. 1999). This size cline may be inter-twined with subspecific differences across the range.Marshall (1967, 1978) recognized no subspecies,but Hekstra (1982) suggested there might be asmany as six subspecies of Flammulated Owl inNorth America (also see McCallum 1994). Morerecently, Pyle (1997) described two subspecies:O. f. idahoensis and O. f. frontalis. Under idahoensis,Pyle suggested that a ‘‘borealis’’ subtype in thenorthern parts of the range might average largerthan its southern counterparts (Pyle 1997).

Flammulated Owls are one of the least sexuallysize-dimorphic North American owls (Earhart andJohnson 1970). It is unknown whether sexual sizedimorphism varies in magnitude across the subspe-cies or latitudinal size cline, or if it is approximatelyconstant across the species’ range, as it is in someother owl species (McGillivray 1989). The difficultyin discriminating females from males also has ob-scured the detection of other common patterns inmigration geography, such as differential migrationin timing or distance by sex (DeLong and Hoffman1999). Although a discriminant function developedto sex owls captured during migration based onexternal measurements was approximately 73% ef-fective (J. DeLong unpubl. data), variation in sizeacross the range suggests that this function may notaccurately identify sex in other populations.

Recent efforts to study the migration ecology ofFlammulated Owls yielded an opportunity to study

the biogeography of size in this species across asoutheast-to-northwest gradient in the western Unit-ed States (DeLong 2004, 2006, DeLong et al. 2005,Stock et al. 2006, Linkhart and Reynolds 2007). Us-ing the combined data from mostly contemporaryautumn banding efforts in New Mexico, Nevada,and Idaho, we present an assessment of sex-, age-,and site-specific morphometric variation in this spe-cies. We use these results to assess differences in sizeand sexual size dimorphism across the sites, and weconsider the implications for understanding thespecies’ migration geography.

METHODS

From 1998–2010, annual capture totals for Flam-mulated Owls banded during autumn migration onBoise Ridge in west-central Idaho ranged from 2–66owls (Stock et al. 2006, Idaho Bird Observatoryunpubl. data: technical reports at http://www.idahobirdobservatory.org). From 1993–2003, au-tumn capture totals in the Manzano Mountains ofcentral New Mexico ranged from 20–156 owls(DeLong 2004, 2006, DeLong et al. 2005; Hawk-Watch International [HWI] unpubl. data: technicalreports at http://www.hawkwatch.org). From 2001–08, autumn capture totals in the Goshute Mountainsof northeastern Nevada ranged from 19–72 owls(HWI unpubl. data: technical reports at http://www.hawkwatch.org). All three study sites are onmountain ridgetops where long-term studies of bothdiurnal and nocturnal raptors occur (Fig. 1).

The Idaho study was in the same area as thelong-term Boise Ridge Raptor Migration Project(43u36.329N, 116u03.629W; elevation ca. 1799 m;Smith et al. 2008); Stock et al. (2006) and Leppertet al. (2006) previously described the study site andrelevant owl capture and banding methods. Mea-surements collected at this site included mass andunflattened wing chord, with standard tail length,exposed culmen length, and tarsus length collectedless regularly. The New Mexico study was in thesame area as the long-term Manzano MountainsRaptor Migration Project (34u42.259N, 106u24.679W;elevation ca. 2805 m; DeLong and Hoffman 1999,Hoffman et al. 2002, Smith et al. 2008); Delonget al. (2005) previously described the owl study siteand relevant methods.

The Nevada study site was in the same area as thelong-term Goshute Mountains Raptor MigrationProject (43u25.429N, 114u16.289W; elevation ca.2740 m; Hoffman 1985, Hoffman et al. 2002,

110 SMITH ET AL. VOL. 46, NO. 1

Smith et al. 2008). The owl capture arena was in amixed conifer forest and sagebrush meadow areaapproximately 150 m below and 150–200 m east ofthe main ridgetop area where most of the diurnalcount and banding operations occurred. The meth-ods used to capture owls were similar to those used inIdaho and New Mexico. During 2008, when samplesfor DNA sexing were collected, banders lured owlsinto a capture arena consisting of 17 stationary mistnets (60-mm mesh, 9–12 m in length) using Flammu-lated Owl calls broadcasted via a recorded audiolurethrough an amplified stereo unit (Reynolds andLinkhart 1984). Weather permitting (absence of pre-cipitation and winds ,24 kph), netting occurrednightly from 23 August through 28 October fromapproximately 30 min after sundown to 30 min be-fore sunrise. Measurements collected included mass,unflattened wing chord, standard tail length, andexposed culmen length, with hallux length and tar-sus length collected less regularly. When possible,banders determined the age of Flammulated Owlsin the field based on flight feather molt and faultbars (Pyle 1997, DeLong 2004). Hatch-year (HY) owls

had retained juvenile plumage or uniform fault bars.After-hatch-year (AHY) owls showed flight-feathermolt or multiple generations of flight feathers.

Sexing Flammulated Owls based on externalcharacteristics is problematic (McCallum 1994). Ineach of the studies considered here, researcherscollected samples to enable DNA-based sexing ofselected owls. Leppert et al. (2006) used DNA tosex 55 owls captured in Idaho, including 16 HY owls(10 female, 6 male), 9 AHY owls (4 females, 5males), and 30 owls of unknown age (9 females,21 males). The New Mexico sample of DNA-sexedowls (DeLong et al. 2005) included 13 female and20 male HY owls, plus another 20 females and 33males of unknown age. The Nevada sample of DNA-sexed owls included 21 female and 37 male HY owls.Wildlife Genetics International (Nelson, British Co-lumbia, Canada) processed all of the New Mexicoand Nevada samples, and all three sets of sampleswere processed following methods developed byGriffiths et al. (1998).

Given availability of DNA sexing data from allthree sites, we initially focused on analyzing siteand sex differences in individual measurementsamong known-sex HY owls. Guided by the resultsof these analyses, we conducted additional analysesthat included other owls of unknown age, sex, orboth. Because not all measurements were taken onall owls, we took a progressive, hierarchical ap-proach to these additional analyses to maximizeconsideration of as many individual owls as possiblewhile investigating sex-, age-, and site-specific varia-tion in individual measurements. Our analyses in-cluded various two-way (site * sex, site * age, sex *age) factorial ANOVAs and, when appropriate, Bon-ferroni post-hoc comparisons to further discern dif-ferences among the three sites (Systat v. 10; SPSS,Inc. 2000). We did not take a multivariate approachwith aggregate measurement indices, such as prin-ciple components analysis, because there were toofew owls from each site with full suites of measure-ments suited to multivariate analysis. In addition,our interest extended beyond examining variationin overall size and shape; we were specifically inter-ested in determining which measurements differedamong sites, sexes, and age classes. Diagnostic plotsconfirmed reasonable conformity to assumptions ofnormally distributed residuals and heteroscedasticvariances. We considered results significant if P #

0.05, but we also highlight marginally significantresults with 0.05 , P # 0.10.

Figure 1. Locations of Flammulated Owl migration studysites in the western United States.

MARCH 2012 FLAMMULATED OWL MORPHOMETRICS 111

Tab

le1.

Co

mp

aris

on

s(2

-way

fact

ori

alA

NO

VA

,le

ast-

squ

ares

mea

ns)

of

the

sex-

spec

ific

mo

rph

om

etri

cso

fD

NA

-sex

ed,

hat

ch-y

ear

Fla

mm

ula

ted

Ow

lsca

ptu

red

du

rin

gau

tum

nm

igra

tio

no

nB

ois

eR

idge

,Id

aho

,in

the

Go

shu

teM

ou

nta

ins,

Nev

ada,

and

inth

eM

anza

no

Mo

un

tain

s,N

ewM

exic

o.

ME

ASU

RE

SIT

E

FE

MA

LE

MA

LE

MO

DE

LR

2SI

TE

PSE

XP

INT.

Pn

ME

AN

SEn

ME

AN

SE

Mas

s(g

)Id

aho

964

.41.

906

63.2

2.32

0.27

3,

0.00

1a0.

432

0.40

7N

evad

a20

60.2

1.27

3760

.90.

94N

ewM

exic

o13

56.3

1.58

2053

.61.

27W

ing

cho

rd(m

m)

Idah

o10

134.

31.

066

132.

21.

370.

291

,0.

001b

,0.

001

0.83

9N

evad

a21

133.

40.

7337

130.

50.

55N

ewM

exic

o13

131.

10.

9320

127.

70.

75E

xpo

sed

culm

enle

ngt

h(m

m)

Idah

o8

8.7

0.17

68.

70.

200.

098

0.06

60.

096

0.73

3N

evad

a21

9.1

0.11

378.

80.

080.

095c

0.06

70.

028

0.61

3N

ewM

exic

o13

8.9

0.14

208.

70.

11H

allu

xle

ngt

h(m

m)

Nev

ada

136.

40.

0831

6.2

0.07

0.13

20.

486

0.00

20.

440

New

Mex

ico

136.

40.

1120

6.0

0.11

Tar

sus

len

gth

(mm

)Id

aho

825

.00.

436

24.3

0.50

0.34

9,

0.00

10.

160

0.83

7N

evad

a7

22.6

0.46

2522

.20.

240.

065c

0.17

00.

372

0.78

6N

ewM

exic

o13

23.0

0.34

2022

.80.

27St

and

ard

tail

len

gth

(mm

)Id

aho

1063

.01.

346

62.0

1.73

0.05

90.

187

0.24

60.

903

Nev

ada

2162

.00.

9336

61.0

0.50

New

Mex

ico

1361

.01.

1820

60.6

0.95

aB

on

ferr

on

ip

airw

ise

com

par

iso

ns,

mo

del

P#

0.05

:Id

aho

5N

evad

a.

New

Mex

ico

.b

Bo

nfe

rro

ni

pai

rwis

eco

mp

aris

on

s,m

od

elP

#0.

05:

Idah

o5

Nev

ada

.N

ewM

exic

o.

cA

nal

ysis

reru

nex

clu

din

gId

aho

dat

a.


RESULTS

Analyses for HY birds of known sex indicatedno significant site * sex interactions; overall sexdifferences (females . males) for wing chord andhallux length; and overall site differences for mass,wing chord, and tarsus length (Table 1, Fig. 2). Asimilar sex difference was indicated for culmenlength, but the difference was only marginally sig-nificant (P 5 0.066); closer inspection revealed

inconsistency across sites, with no difference insex-specific means for culmen in Idaho. Site-specificmeans revealed a clinal pattern of declining mass,wing chord, tail length, and hallux length (NV toNM only) from northwest to southeast, with post-hoc comparisons confirming significant differencesbetween Idaho and New Mexico for mass and wingchord. A marginally significant (P 5 0.096) site dif-ference also was indicated for culmen length;

Figure 2. Box and whisker plots (Systat 10.0, SPSS Inc. 2000) illustrating variation by sex (DNA-based) and site (fromleft to right within sexes: Boise Ridge, ID, Goshute Mountains, NV, and Manzano Mountains, NM) in measurements ofhatch-year Flammulated Owls captured during autumn migration in the western United States.


Tab

le2.

Co

mp

aris

on

s(2

-way

fact

ori

alA

NO

VA

,le

ast-

squ

ares

mea

ns)

of

the

age-

spec

ific

mo

rph

om

etri

cso

fF

lam

mu

late

dO

wls

(fem

ale,

mal

e,an

du

nkn

ow

n-s

exco

mb

ined

)ca

ptu

red

du

rin

gau

tum

nm

igra

tio

no

nB

ois

eR

idge

,Id

aho

,in

the

Go

shu

teM

ou

nta

ins,

Nev

ada,

and

inth

eM

anza

no

Mo

un

tain

s,N

ewM

exic

o.

ME

ASU

RE

SIT

E

HA

TC

H-Y

EA

RA

FT

ER-H

AT

CH

-YE

AR

MO

DE

LR

2SI

TE

PA

GE

PIN

T.

Pn

ME

AN

SEn

ME

AN

SE

Mas

s(g

)Id

aho

1563

.91.

329

63.6

1.70

0.32

1,

0.00

1a0.

043

0.06

2N

evad

a15

159

.20.

4239

60.6

0.82

New

Mex

ico

168

53.1

0.39

2357

.41.

06W

ing

cho

rd(m

m)b

Idah

o16

133.

50.

939

134.

11.

240.

214

,0.

001c

0.00

40.

546

Nev

ada

152

131.

80.

3039

134.

20.

59N

ewM

exic

o16

812

8.8

0.29

2313

1.2

0.77

Exp

ose

dcu

lmen

len

gth

(mm

)bId

aho

148.

70.

179

8.5

0.21

0.10

5,

0.00

10.

190

0.10

7N

evad

a14

09.

10.

0537

9.4

0.10

0.08

1d0.

020

,0.

001

0.47

7N

ewM

exic

o16

68.

90.

0523

9.3

0.13

Hal

lux

len

gth

(mm

)bN

evad

a10

46.

20.

0532

6.5

0.08

0.05

20.

190

0.00

60.

473

New

Mex

ico

110

6.2

0.05

136.

30.

14T

arsu

sle

ngt

h(m

m)

Idah

o14

24.7

0.38

924

.70.

470.

143

,0.

001

0.43

60.

762

Nev

ada

6422

.80.

1818

23.3

0.33

0.00

8d0.

801

0.30

40.

550

New

Mex

ico

112

22.9

0.13

1323

.10.

39St

and

ard

tail

len

gth

(mm

)Id

aho

1662

.60.

829

62.1

1.09

0.02

40.

131

0.84

30.

712

Nev

ada

150

61.5

0.23

3861

.40.

53N

ewM

exic

o16

760

.80.

2523

61.2

0.68

aB

on

ferr

on

ip

airw

ise

com

par

iso

ns,

mo

del

P#

0.05

:Id

aho

.N

evad

a.

New

Mex

ico

.b

No

teth

atth

ese

anal

yses

may

be

con

fou

nd

edb

ysi

gnif

ican

tse

x-sp

ecif

icva

riat

ion

(see

Tab

le1)

,bu

tw

ere

con

du

cted

toen

able

ten

tati

vein

vest

igat

ion

so

fag

e-an

dsi

te-s

pec

ific

vari

atio

nb

ased

on

larg

ern

on

-sex

-sp

ecif

icsa

mp

lesi

zes.

cB

on

ferr

on

ip

airw

ise

com

par

iso

ns,

mo

del

P#

0.05

:Id

aho

and

Nev

ada

.N

ewM

exic

o.

dA

nal

ysis

reru

nex

clu

din

gId

aho

dat

a.


however, although the pattern of increasing mea-surements held between New Mexico and Nevada,unlike all other measurements, culmen lengths av-eraged shortest in Idaho. Tarsi averaged significant-ly longer in Idaho than in Nevada and New Mexico;however, unlike for any other measurement, both

sexes averaged nonsignificantly shorter tarsi in Nevadathan in New Mexico (Table 1). Noting these results forIdaho as potentially anomalous (see Discussion), weran the analyses for culmen and tarsus length againexcluding Idaho data. This yielded no site or sexdifferences for tarsus length, but a significant sex

Figure 3. Box and whisker plots (Systat 10.0, SPSS Inc. 2000) illustrating variation by age (AHY 5 after-hatch-year, HY 5

hatch-year) and site (from left to right within sexes: Boise Ridge, ID, Goshute Mountains, NV, and Manzano Mountains,NM) in measurements of Flammulated Owls (sexes combined) captured during autumn migration in the westernUnited States.


Tab

le3.

Co

mp

aris

on

s(2

-way

fact

ori

alA

NO

VA

,lea

st-s

qu

are

mea

ns)

of

the

sex-

spec

ific

mo

rph

om

etri

cso

fF

lam

mu

late

dO

wls

(HY,

AH

Y,an

du

nkn

ow

n-a

geco

mb

ined

)ca

ptu

red

du

rin

gau

tum

nm

igra

tio

no

nB

ois

eR

idge

,Id

aho

,in

the

Go

shu

teM

ou

nta

ins,

Nev

ada,

and

inth

eM

anza

no

Mo

un

tain

s,N

ewM

exic

o.

ME

ASU

RE

SIT

E

FE

MA

LE

MA

LE

MO

DE

LR

2SI

TE

PSE

XP

INT.

Pn

ME

AN

SEn

ME

AN

SE

Mas

s(g

)aId

aho

2262

.81.

0732

63.6

0.89

0.35

4,

0.00

1b0.

770

0.52

4N

evad

a20

60.2

1.13

3760

.90.

83N

ewM

exic

o32

55.3

0.89

4254

.40.

78W

ing

cho

rd(m

m)a

Idah

o23

134.

30.

6932

133.

70.

590.

304

,0.

001b

,0.

001

0.21

0N

evad

a21

133.

40.

7237

130.

50.

54N

ewM

exic

o33

130.

30.

5842

128.

50.

51E

xpo

sed

culm

enle

ngt

h(m

m)

Idah

o18

8.6

0.12

188.

60.

120.

104

0.00

30.

024

0.43

9N

evad

a21

9.1

0.11

378.

80.

080.

100c

0.00

90.

005

0.72

0N

ewM

exic

o33

8.8

0.09

428.

60.

08H

allu

xle

ngt

h(m

m)a

Nev

ada

136.

40.

1131

6.2

0.07

0.05

80.

112

0.02

00.

538

New

Mex

ico

326.

20.

0742

6.1

0.06

Tar

sus

len

gth

(mm

)Id

aho

1824

.80.

2818

24.8

0.28

0.39

6,

0.00

1d0.

328

0.86

5N

evad

a7

22.6

0.45

2522

.20.

240.

046c

0.16

80.

321

0.72

1N

ewM

exic

o33

22.9

0.21

4222

.70.

19St

and

ard

tail

len

gth

(mm

)Id

aho

2362

.40.

6020

62.7

0.65

0.07

50.

005

0.35

80.

575

Nev

ada

2162

.00.

6336

61.0

0.48

New

Mex

ico

3361

.00.

5033

60.5

0.45

aT

hes

ean

alys

esm

ayb

eco

nfo

un

ded

by

sign

ific

ant

age-

spec

ific

vari

atio

n(s

eeT

able

2),b

ut

wer

eco

nd

uct

edto

enab

lete

nta

tive

inve

stig

atio

ns

of

sex-

and

site

-sp

ecif

icva

riat

ion

bas

edo

nla

rger

no

n-a

ge-s

pec

ific

sam

ple

size

s.b

Bo

nfe

rro

ni

pai

rwis

eco

mp

aris

on

s,m

od

elP

#0.

05:

Idah

o.

Nev

ada

.N

ewM

exic

o.

cA

nal

ysis

reru

nex

clu

din

gId

aho

dat

a.d

Bo

nfe

rro

ni

pai

rwis

eco

mp

aris

on

s,m

od

elP

$0.

05:

Idah

o.

Nev

ada

and

New

Mex

ico

.


difference (female . male) and marginally significantsite difference (Nevada . New Mexico; P 5 0.067) forculmen length (Table 1).

With no DNA-based sexing data for known AHY Ne-vada or New Mexico owls, we were restricted to exam-ining the Idaho data for possible age-related variation inwing chord and culmen lengths, while simultaneouslyaccounting for apparent sex effects. This analysis re-vealed only marginally (P 5 0.094) longer wing chordsfor females (least-squares mean 6 SE: 134.9 6 0.92)compared with males (132.6 6 0.94); however, smallsample sizes (n 5 6 HY male, 5 AHY male, 10 HYfemale, and 4 AHY female) compromised the statisticalpower of this analysis. Also note that, in Idaho, AHY owlsaveraged nonsignificantly lighter (63.5 6 2.00 vs. 63.8 6

1.58 g), longer wing chords (134.3 6 1.05 vs. 133.2 6

0.81 mm) and tarsi (24.7 6 0.26 vs. 24.6 6 0.21 mm),and shorter culmens (8.5 6 0.16 vs. 8.7 6 0.13 mm) andtails (62.1 6 0.91 vs. 62.5 6 0.70 mm) than HY owls.

Because there was no evidence of sex-specific dif-ferences, we expanded the analyses of mass, taillength, and tarsus length to include owls of knownage but unknown sex from Nevada and Idaho(Table 2, Fig. 3). Analysis of mass revealed signifi-cant site and age main effects and a marginallysignificant sex * age interaction. Post-hoc compari-sons confirmed Idaho . Nevada . New Mexico.AHY owls tended to average greater mass than HYowls. The marginally significant interaction reflect-ed this tendency diminishing from southeast tonorthwest, with no age-related mass difference ap-parent in Idaho. Analysis of tarsus length, excludingIdaho data, revealed no site or age effects and noconsistent pattern of difference between Nevadaand New Mexico within age groups. Analysis of taillengths revealed no significant effects, but againsuggested at least a slight declining trend fromnorthwest to southeast.

To enable larger-sample site comparisons we ex-panded analyses of wing chord, culmen length, andhallux length to include owls of known age butunknown sex (Table 2, Fig. 3), keeping in mindthat sex-specific differences may confound thesecomparisons. AHY owls averaged longer wingchords and halluxes than HY owls, and wing chordsaveraged shorter in New Mexico than in Nevada andIdaho. Culmen lengths averaged longer amongAHY birds in both Nevada and New Mexico, butnot in Idaho. With Idaho data excluded, culmensaveraged longer in Nevada than in New Mexico; asimilar but nonsignificant (P . 0.10) difference wasshown for hallux length.

Because there was no evidence of age-specific dif-ferences in either tarsus or tail length, we expandedanalyses of these measurements to include owls ofknown sex but unknown age from Idaho and NewMexico (Table 3). Similar to the HY-only analyses,the expanded analyses indicated no significant sexdifferences for either measurement and no signifi-cant difference in tarsus length between Nevadaand New Mexico. Unlike the HY-only analysis, how-ever, the new analysis indicated longer tails in Idahothan in New Mexico, with Nevada not significantlydifferent from either. Given no evidence of sex orage differences, final site-only analyses incorporatingall available data (including additional owls of un-known age and sex from all sites, but primarily Ne-vada and Idaho) again confirmed no difference intarsus lengths between Nevada (least-squares mean6 SE: 22.9 6 0.145 mm) and New Mexico (22.8 6

0.07 mm; r2 5 0.003, F1,457 5 1.17, P 5 0.280), and adistinct clinal pattern of increase in tail length fromsoutheast to northwest (NM: 61.1 6 0.13, n 5 453;NV: 61.3 6 0.17, n 5 249; ID: 62.6 6 0.40, n 5 47). Inthis case, however, the tails of Idaho owls emerged assignificantly longer than those of Nevada and NewMexico owls (r2 5 0.017, F2,749 5 6.35, P 5 0.002).

To produce more robust sex and site compari-sons, we expanded analyses of mass, wing chord,culmen length, and hallux length to include owlsof known sex but unknown age (Table 3), keepingin mind that age differences may confound thesecomparisons. These expanded analyses again con-firmed overall sex differences (female . male) forwing chord, culmen length, and hallux length, andsite differences for mass, wing chord, and culmenlength (NV . NM; ID excluded). Post-hoc compari-sons confirmed Idaho . Nevada . New Mexico formass and wing chord.

DISCUSSION

Geographic variation in size is of broad interest toecologists and evolutionary biologists (Brown andLomolino 1998). For some species, including theFlammulated Owl, biogeographic patterns havebeen difficult to document due to the difficulty ofobtaining geographically distributed samples. Wecontribute to overcoming this historical limitationby combining data from three studies distributedacross the western United States to identify patternsof size variation among sites, ages, and sexes.

Most measurements, especially mass and wingchord, showed evidence of an increase along a south-east-to-northwest gradient. These results confirm the


latitudinal cline suggested by Marshall (1967) andare consistent with Pyle’s (1997) subspecies classifi-cation. Interestingly, two anomalies arose: (a) cul-mens averaged longest in Nevada and shortest inIdaho; and (b) tarsi averaged longer in Idaho, butno significant or consistent (within sexes or agegroups) differences were apparent in comparingowls from Nevada and New Mexico. Thus, the clinemay not be consistent across all measurements; how-ever, whether these anomalies reflect ecologicalfactors that differ among sites, sampling errors, orperhaps the influence of relatively small sample sizesin Idaho, remains unknown. For example, longerwings and tails may correlate with northern birdsmigrating longer distances, and longer culmensmay correlate with northern birds taking larger in-sect prey. Skeletal measurements such as tarsuslength often are the most reliable for detectingchanges in overall size (Smith 1988), but this mea-surement was not taken at the Idaho site.

Within sites, females averaged longer wing chordsand, to a lesser degree, longer culmens (except inIdaho) and halluxes than males, with similar butnonsignificant differences usually shown for othermeasurements. Thus, a very slight reversed sexualsize-dimorphism was evident, which remained rough-ly constant across sites (Table 3). Therefore, it is un-likely that a discriminant function to identify sex inthese owls would function well from one site to thenext, although it remains technically plausible thatfurther research focused on a geographically restrict-ed breeding population could enable reliable dis-crimination of sex from external morphometrics.Although statistically significant, the magnitude ofmany differences was small; i.e., a few grams/milli-meters or ,5% difference. Nevertheless, some site-specific differences may reflect important elementsof sex-specific ecology. For example, similar mass butlonger wings suggests lighter wing loading in fe-males. Though difficult to investigate in a nearlymonomorphic species, this characteristic may sug-gest that females migrate longer distances to moresoutherly winter ranges than males. However, theconsistent degree of sexual size dimorphism suggeststhe possibility that northern birds do not necessarilytravel farther than southern birds. In addition, lon-ger culmen and hallux measurements in females maysuggest the selection of larger prey compared withmales, possibly resulting in a broader combined foodniche for these insectivorous owls.

Within sites, AHY owls usually averaged heavier andlonger wing chords, culmens, and halluxes than HY

owls (Table 2). The difference in mass and culmenlength diminished from southeast to northwest, how-ever, with no age difference evident in Idaho for thesemeasurements. This mass result suggests differencesin body condition and migratory activity among thesites. In New Mexico, HY birds had lower fat and mus-cle content than adults, but HY birds gradually gotheavier as the migration season progressed (DeLonget al. 2005). The presence of some owls at the sitethroughout the season suggested that many HY owlsat the New Mexico site were in a premigratory phase,departing on migration only after their body condi-tion was adequate (DeLong et al. 2005). The fact thatHY and AHY owls in Nevada and Idaho showed littledifference in mass suggested that either HY owls hadalready attained adult mass and were actively migrat-ing or AHY owls were in poor condition relative to HYowls. Because energy reserves are a crucial aspect ofthe migratory ecology of most birds (Bairlein andGwinner 1994), such differences may signify geograph-ic variation in migratory strategy within this species.

It appeared anomalous that culmen length aver-aged shortest in Idaho rather than extending themore typical pattern of an increasing cline in sizefrom southeast to northwest. Similarly, it appearedanomalous that tarsus lengths averaged substantiallylonger in Idaho than in both Nevada and New Mex-ico, but no significant or consistent differences wereevident between the latter two sites. This contrastedwith the clinal patterns shown for most other mea-surements where Idaho . Nevada . New Mexicoor Idaho/Nevada . New Mexico. These two anoma-lies suggested possible biases due to either differentmeasuring protocols, with the Nevada and New Mex-ico data collected by one organization and the Idahodata by another, or the influence of lower samplesizes in Idaho. Eliminating this concern from furtherconsideration by reducing the comparison to Nevadaand New Mexico led to more robust final conclusions:(1) no significant variation in tarsus length due to sex,age, or site; and (2) significant variation in culmenlength due to sex (longer in females), age (longer inAHY owls), and site (longer in NV than in NM).

Unless migrating birds travel well east of their sum-mer ranges in Idaho and the Pacific Northwest to flydown through the Rocky Mountains of New Mexico,larger birds should be more common among the Ida-ho-Nevada as opposed to New Mexico migrants, be-cause the Boise Ridge and Goshute Mountains sitesboth lie several hundred kilometers farther north andlikely draw migrants from a broader range of northerngeography. Put another way, significant differences in


the average size of Idaho-Nevada and New Mexicomigrants clearly suggested that larger northern birdsare an uncommon component of the New Mexicomigrant population. This notion also was consistentwith evidence from analysis of hydrogen stable-isotoperatios in feathers of New Mexico owls, which suggestedthat most owls captured at this site originated relativelynearby in Colorado and northern New Mexico (De-Long et al. 2005). In addition, although less frequentlysignificant, evidence of detectable differences in thesize ranges of Nevada and Idaho owls further suggest-ed relatively limited mixing of the populations thatmigrate through these two sites.

Variable sample sizes and incomplete representa-tion across the three study sites of accurately sexedand aged owls complicated this investigation. Never-theless, a careful series of hierarchical analyses, gradu-ally incorporating data for additional owls of unknownage, sex, or both, yielded valuable new insights intoboth sex- and age-specific variation, as well as regionalvariation in the morphometrics of Flammulated Owls.

ACKNOWLEDGMENTS

Funding for the Nevada study was provided by the Bu-reau of Land Management–Elko District Office and HWImembers. HWI extends special thanks to Chris Neri forinitiating the Goshute owl project. Other key field assis-tants included Dan Brown, Alicia Byrd, Mary Cheney, Sa-rah Faegre, Anna Knipps, Dawn Konkoly, Nova McKentley,Alberto Martinez, James Melton, Dave Roth, and MeganSchwender. Funding for the New Mexico study was provid-ed by the USDA Forest Service–Cibola National Forest andRegion 3, New Mexico Game and Fish Department–Sharewith Wildlife Program, Public Service Company of NewMexico, New Mexico Ornithological Society, Gayle andJim Higman, and Jeanie and Jodie Humber. Key field as-sistants included Wendy Beard, Bruce Casler, Anne Con-don, Sarah Hamilton, Zach Hurst, Wendy King, TraceyMader, and Elise Apple Snider. Funding for the Idahostudy was provided by the USDA Forest Service–Boise Na-tional Forest, USDI Fish and Wildlife Service–Boise FieldOffice, Idaho Department of Fish and Game, Raptor Re-search Center at Boise State University, Idaho Chapter ofthe Wildlife Society, Association of Field Ornithologists,Wilson Ornithological Society, Potlatch Corporation, Ma-zamas, Boise Cascade, Sigma Xi, and IBO private donors.Key field assistants included Dan Battaglia, Julia Camp,Tami Denette, Ben Flemer, Greg Greene, Frank Mayer,Anaka Mines, Matthew Moskwik, and Kelly Ward.

LITERATURE CITED

BAIRLEIN, F. AND E. GWINNER. 1994. Nutritional mechanismsand temporal control of migratory energy accumulationin birds. Annual Review of Nutrition 14:187–215.

BALDA, R.P., B.C. MCKNIGHT, AND C.D. JOHNSON. 1975.Flammulated Owl migration in the southwestern Unit-ed States. Wilson Bulletin 87:520–533.

BANKS, R.C. 1964. An experiment on a Flammulated Owl.Condor 66:79.

BLACKBURN, T.M., K.J. GASTON, AND N. LODER. 1999. Geo-graphic gradients in body size: a clarification of Berg-mann’s rule. Diversity and Distributions 5:165–174.

BROWN, J.H. AND M.V. LOMOLINO. 1998. Biogeography. Si-nauer Associates, Inc., Sunderland, MA U.S.A.

DELONG, J.P. 2004. Age determination and preformativemolt in hatch-year Flammulated Owls during the fall.North American Bird Bander 29:111–115.

———. 2006. Pre-migratory fattening and mass gain inFlammulated Owls in central New Mexico. Wilson Jour-nal of Ornithology 118:187–193.

——— AND S.W. HOFFMAN. 1999. Differential autumn mi-gration of Sharp-shinned and Cooper’s hawks in west-ern North America. Condor 101:674–678.

———, T. MEEHAN, AND R. SMITH. 2005. Investigating fallmovements of Flammulated Owls (Otus flammeolus) incentral New Mexico using stable hydrogen isotopes.Journal of Raptor Research 39:19–25.

EARHART, C.M. AND N.K. JOHNSON. 1970. Size dimorphismand food habits of North American owls. Condor72:251–264.

GRIFFITHS, R., M.C. DOUBLE, K. ORR, AND R.L.G. DAWSON.1998. A DNA to sex most birds. Molecular Ecology7:1071–1075.

HEKSTRA, G.P. 1982. Description of twenty-four new sub-species of American Otus. Bulletin of the Zoological Mu-seum of the University of Amsterdam 9:49–63.

HOFFMAN, S.W. 1985. Autumn Cooper’s Hawk migrationthrough northern Utah and northeastern Nevada,1977–1982. Pages 149–165 in M. Harwood [ED.], Pro-ceedings of Hawk Migration Conference IV. Hawk Mi-gration Association of North America.

———, J.P. SMITH, AND T.D. MEEHAN. 2002. Breedinggrounds, winter ranges, and migratory routes of raptorsin the mountain west. Journal of Raptor Research36:97–110.

LEPPERT, L., T.V. ZADOROZHNY, J.R. BELTHOFF, A.M. DUFTY,JR., S.L. STOCK, G. KALTENECKER, AND J.E. SMITH. 2006.Sex identification of four owl species from Idaho:DNA and morphometrics. Journal of Raptor Research10:291–294.

LIGON, J.D. 1968. Starvation of spring migrants in the Chir-icahua Mountains, Arizona. Condor 70:387–388.

LINKHART, B.D. AND R.T. REYNOLDS. 2007. Return rate,fidelity, and dispersal in a breeding population ofFlammulated Owls (Otus flammeolus). Auk 124:264–275.

MARSHALL, J.T. 1967. Parallel variation in North and Mid-dle American screech-owls. Western Foundation of Verte-brate Zoology 1:1–72.

———. 1978. Systematics of smaller Asian night birdsbased on voice. Ornithological Monographs 25:1–58.

MCCALLUM, D.A. 1994. Flammulated Owl (Otus flammeo-lus). In A. Poole and F. Gill [EDS.], The birds of NorthAmerica, No. 93. The Academy of Natural Sciences,


Philadelphia, PA and The American Ornithologists’Union, Washington, DC U.S.A.

MCGILLIVRAY, W.B. 1989. Geographic variation in size andreverse size dimorphism of the Great Horned Owl inNorth America. Condor 91:777–786.

PHILLIPS, A.R. 1942. Notes on the migration of Elf andFlammulated Screech owls. Wilson Bulletin 54:132–137.

PYLE, P. 1997. Identification guide to North Americanbirds, Part I. Slate Creek Press, Bolinas, CA U.S.A.

REYNOLDS, R.T. AND B.D. LINKHART. 1984. Method and ma-terials for capturing and monitoring FlammulatedOwls. Great Basin Naturalist 44:49–51.

RICH, T.D., C.J. BEARDMORE, H. BERLANGA, P.J. BLANCHER,M.S.W. BRADSTREET, G.S. BUTCHER, D.W. DEMAREST, E.H.DUNN, W.C. HUNTER, E.E. INIGO-ELIAS, J.A. KENNEDY, A.M.MARTELL, A.O. PANJABI, D.N. PASHLEY, K.V. ROSENBERG,C.M. RUSTAY, J.S. WENDT, AND T.C. WILL. 2004. Partnersin Flight North American landbird conservation plan. Cor-nell Lab of Ornithology, Ithaca, NY U.S.A. http://www.partnersinflight.org/cont_plan (last accessed August 2011).

SMITH, J.P. 1988. Morphometric variation in accipiterhawks with emphasis on western North America. M.S.thesis. Utah State Univ., Logan, UT U.S.A.

———, C.J. FARMER, S.W. HOFFMAN, K.Z. WOODRUFF, G.S.KALTENECKER, AND P. SHERRINGTON. 2008. Trends inautumn counts of migratory raptors in western NorthAmerica. Pages 217–251 in K.L. Bildstein, J.P. Smith, E.Ruelas Inzunza, and R.R. Veit [EDS.], State of NorthAmerica’s birds of prey. Series in Ornithology 3. NuttallOrnithological Club, Cambridge, MA and AmericanOrnithologists’ Union, Washington, DC U.S.A.

STOCK, S.L., P.J. HEGLUND, G.S. KALTENECKER, J.D. CAR-

LISLE, AND L. LEPPERT. 2006. Comparative ecology ofthe Flammulated Owl and Northern Saw-whet Owl dur-ing fall migration. Journal of Raptor Research 40:120–129.

WINTER, J. 1974. The distribution of the Flammulated Owlin California. Western Birds 5:25–44.

Received 15 November 2010; accepted 18 October 2011Associate Editor: Bruce Peterjohn


during Autumn Migration in the Western United States...

Documents

Transcript of during Autumn Migration in the Western United States...