Morphometric Variation in Marbled Murrelets, Brachyramphus Marmoratus, in British
Columbia
Author(s): Cindy L. Hull, Brett A. Vanderkist, Lynn W. Lougheed, Gary W. Kaiser and Fred
Cooke
Source: Northwestern Naturalist, Vol. 82, No. 2 (Autumn, 2001), pp. 41-51
Published by: Society for Northwestern Vertebrate Biology
Stable URL: http://www.jstor.org/stable/3536785
Accessed: 27-05-2015 17:51 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
Society for Northwestern Vertebrate Biology is collaborating with JSTOR to digitize, preserve and extend access to
Northwestern Naturalist.
http://www.jstor.org
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
NORTHWESTERN
NATURALIST
82:41-51
AUTUMN 2001
MORPHOMETRIC VARIATION IN MARBLED MURRELETS,
BRACHYRAMPHUS MARMORATUS, IN BRITISH COLUMBIA
CINDY L. HULL1, BRETT A. VANDERKIST,AND LYNN W. LOUGHEED2
CWS/NSERC WildlifeEcologyResearchChair,Department
BiologicalSciences,
8888 University
SimonFraserUniversity,
Drive,Burnaby,BC V5A 1S6 CANADA
GARY W. KAISER
Canadian WildlifeServiceRR1 5421 RobertsonRoad,Delta, BC V4K 3N2 CANADA
FRED COOKE
CWS/NSERC WildlifeEcologyResearchChair,Department
BiologicalSciences,
8888 University
SimonFraserUniversity,
Drive,Burnaby,BC V5A 1S6 CANADA
ABSTRACT-Morphometrics(culmen length,bill height,bill width, wing chord length,and
tarsus length) were taken on 664 marbled murreletsat Desolation Sound and Mussel Inlet,British Columbia, during 1994 to 1997, in order to assess morphological differenceswithin and
between populations and the accuracy of a discriminantfunctionanalysis to identifythe sex of
birds. An assessment of inter-and intra-observervariabilityin measurements was also made.
Significantinter-observereffectsand some intra-observereffectswere found. Data fromrecaptured murreletsindicated they had decreased in size with age, which was attributedto interobserver effects.Wing chord lengthhad the highest measurementerror(66.8%) among observers and tarsus lengthhad the lowest error(36.8%). Deviations of measurementsfromthe mean
were compared among years and sites. No inter-annualdifferenceswere detected in any morphometricat Desolation Sound. Significantdifferencesin culmen length,wing chordlength,and
tarsus lengthwere found between birds fromMussel Inlet and Desolation Sound, which might
indicate discrete populations. The degree of sexual dimorphismin this species was small (measurementsof females average 98% of correspondingmeasurementsof males) and discriminant
functionanalysis revealed only about a 70% success rate in allocating birds to sex; therefore,its
widespread use in this species is not recommended. Future studies of marbled murrelets,or
other avian species, which involve large numbers of personnel, should incorporate extensive
trainingof all observers,with data continually cross-checkedin order to minimize intra-and
inter-observerdifferencesin measurements.
Key words: marbled murrelet, Brachyramphusmarmoratus,morphometrics, Desolation
Sound, Mussel Inlet, BritishColumbia
Morphological measurements of birds are a
valuable means of describing differences among
individuals, cohorts, populations, and closely related species (Barrowclough 1992; Dzubin and
Cooch 1992). They can also be used to accurately
and reliably identify the sex of individuals with
monomorphic plumage (for example, Counsil1 Present address: Environmental
Services,HydroTas-
man and others 1994; Koffijbergand van Eerden
1995; Phillips and Furness 1997; Hedd and others
1998). However, for some avian species, morphological descriptions of populations are limited, as
is an assessment of the use of morphometrics to
identify the sex of birds in the field.
Marbled murrelets, Brachyramphusmarmoratus, are small alcids (family Alcidae) found
along the Pacificcoast of North America from
northern California to Alaska. To date only 2
Tasmania
mania,GPO Box355,Hobart,
7001,AUSTRAstudies have described the morphometrics of
LIA.
2 Present address: Estacian CientificaCharlesDarwin, this species, 1 at Langara Island (Queen Charlotte Islands), British Columbia, Canada, and
PO Box17-01-3891,
Quito,EQLADOR.
41
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
42
NORTHWESTERN NATURALIST
the otherin Alaska (Sealy 1975; Pitocchelliand
others 1995); therefore,little is known of the
discretenessof populations withinthe range of
the species. Further,like most alcids, marbled
murrelets are monomorphic in plumage and
cannot
body size (Bedard 1985) and, therefore,
be sexed in the hand. Although a reliable molecular technique forsexing them has recently
been established (Vanderkistand others 1999),
it necessitates taking a sample of blood and
having expertise in DNA laboratory techniques. Therefore,it would be advantageous if
a technique for reliably and accurately determiningthe sex ofmarbled murreletsin thefield
was established. While both previous studies
described morphological differencesbetween
the sexes of murrelets,neitherassessed the accuracy and reliabilityof using morphometric
charactersto identifythe sex of individuals.
Beforea discriminantfunctionfordetermining the sex of individuals of a species can be
used, the extentof variationwithin the species
needs to be known,as do thebiases fromintraand inter-observerinconsistencies (Lougheed
and others 1991; Phillips and Furness 1997).
Few seabird studies have assessed the effectof
inter-observer differences in measurement
techniques (Barrettand others 1989), although
the use of a number of observers is oftennecessary in many longer-termstudies.
The aims of this study were to: (1) describe
aspects of the morphology of marbled murrelets from2 sites in coastal BritishColumbia and
compare the discreteness of these populations
to data fromotherpopulations, (2) assess morphological variation of murreletsamong years
and times of day (to determineif morphologically dissimilar parts of the population use the
inlets at differenttimes of day), (3) determine
ifmorphometriccharacterscould be used to accuratelyidentifythe sex of marbled murrelets,
and (4) assess the degree of inter-and intra-observervariabilityin measurementsthroughout
the study.
82(2)
1995; Kaiser and others 1995; Derocher and
others 1996) and by 'night-lighting'(modified
from Whitworthand others 1997). Mist nets
were operated for2 to 3 hr around dusk (2000
to 2300, local time) and at dawn (0400 to 0700),
with birds being caught as theymoved to and
from forest areas, presumably on foraging
trips. Night-lightingwas carried out in the
open waters of Desolation Sound from midMay to early August 1997 and in Mussel Inlet
fromJune24 to July9, 1997. Birds were caught
at night (2300 to 0500) when theywere on the
water, by searching in boats using spotlights.
Once a bird was spotted it was approached
slowly and scooped into the boat using a landing net. Night-lightingallowed the capture of
breeding birds earlier in the season than did
mist-netting and the capture of recently
fledged juveniles forthe 1sttime at Desolation
Sound.
Captured birds were weighed and banded
with stainless steel Canadian WildlifeService/
U.S. Fish and Wildlife Service bands. The following morphometricswere taken to the nearest 0.1 mm, using calipers except where stated
(based on Prater and others 1977): flattened
wing chord length (measured from the wrist
joint to the tip of the longest primaryusing a
metal ruler), length of exposed culmen, bill
width (measured just anteriorof thenares), bill
height (measured just anteriorto the edge of
the cuticle), and tarsus length (measured from
theproximal end ofthe tarsus at theinter-tarsal
joint to the footscale at which the toes emerge).
No tarsus measurements were taken at Desolation Sound during 1995). We reportmorphometricdetails as mean I standard deviation.
Two ml of blood were taken fromthe tarsal
vein to determinethe sex of individuals by using the molecular technique of Vanderkistand
others (1999). The birds were then released.
Inter-observer
Effects
More than 20 people measured marbled
murreletsduring thisstudy.The degree ofconSTUDY AREA AND METHODS
sistency in measurements among observers
Field work in BritishColumbia was under- (Barrettand others 1989; Hamer and Furness
taken at Desolation Sound (50004'N, 124?42'W) 1991) was assessed withineach year.Data were
fromMay to August, 1994 to 1997, and at Mus- used fromobserverswho had measured a minsel Inlet (52051'N, 128?10'W, 380 km north of imum of 6 murrelets;because therewas only 1
Desolation Sound) fromJune24 to July9, 1997. measurer in 1994, within-year comparisons
Marbled murreletswere captured using a shal- were only made using data from1995 to 1997.
low-water mist net system (Burns and others Two observersworked on the program for>1
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
AUTUMN
2001
HULL AND OTHERS: MARBLED
yr and theirdata were compared among years
in an effortto ascertain whether inter-annual
variabilityin marbled murreletswas real or as
a result of inter-observereffects.
The morphometricsof marbled murreletsrecaptured during the study period were also assessed to determine if morphological changes
occurred over time, and to get an estimate of
percent measurement error (%ME) for each
morphologicalcharacter(see Bailey and Byrnes
1990). Because only 1 murreletwas recaptured
>1 time,comparisons were made between the
original capture year and the subsequent recapture year.PercentME was calculated by determiningthe within and among bird components of variance and then using a formula
based on Bailey and Byrnes (1990) (but, as recaptured birds were not always measured by
the same observer,this is an adaptation of the
Bailey and Byrnesmodel):
%ME = [Swithin2/Samong2+ Swithin2)] X 100
Comparisons
Morphometric
Adult and recentlyfledged juvenile marbled
murreletswere assessed separately,and only
birds thathad been sexed using the molecular
technique (Vanderkist and others 1999) were
used in the analyses. Juveniles were determined by plumage (Carter and Stein 1995) and
the presence of egg teeth (during this study
some marbled murreletswere observed to have
2 egg teeth).
In order to examine inter-and intra-population variation, comparisons were made between adults and juveniles (the latterwere only
caught in 1997), sexes, years, sites (Desolation
Sound versus Mussel Inlet),methods (mist-netting versus night-lighting),and capture sessions (morning, evening, and all night). Data
fromthe most experienced observers with the
largest sample sizes were used in the analyses.
Comparisons of culmen lengthand wing chord
lengthamong years, sexes, sites, methods,and
sessions were made using ANOVAs and t-tests.
Due to lack of normalityin the data thattransformationsdid not resolve, the othervariables
(bill width, bill height,and tarsus length)were
compared using non-parametric Mann-Whitney and Kruskal-Wallis tests. In order to remove inter-observer effects when assessing inter-annual and site differences, the deviation of
each measurement from the mean of all mea-
43
MORPHOMETRICS
MURRELET
a)
.0
CO
00(,
i.O-t0m0
0,-
\.6d
r;dd\dK116
NN~
~N
u
0
-4
L~
~
CO Cl CO
cu
NN
N
'-0
cO CO CO COCOC
0DmON 0 iU)
UU?
Lf
\6
16
L?
L?
\6
Ln
f?
L6
L6
cU)
N
C7\
00
a,\
CD
o\
nM~~C)
Ln
M
\X)r
cLn
en
O n
000000000000
)
\C0
74
CO
in in in 10 in 1
)
i~ CO~~
D
00 \o
0 .
(6r-
~ in
NN
000N00
CC?
Lf?C?
Ln
O
mCD
l\00N
mCOC
4-S
ON \L
\0an0
0N
ON
. OCD
0 CD
0 0
+ 1
+1 +1 +1 +1 +1 +1+
1 +1 1+1+1
*20
?c/
.55
ZN
0o \0O10 in
~Nl
.0,
0.0
~
~ CL~
r
00
O
re to \O .e
+1 +1 +1 +1 +1 +1 +1 + +1 +1 +1+1
CO
0
0CI e
E; a
t
COO00
0t
0ON
S
-4
00xNCl
N
-4*
Ox 00 0N
ON N N
O tO CO
00 00
Ne
.
0000C0NN0\Co
+1 +1 +1 +1 +1 +1 +1+1
+1 +1
4-j
N I
N
00\ ON000
0
ON0 0\
ClN
Q)
0)
?1
SOn
NNo
0N0NN00
NmC) i CO
O 0C
O
rbOC
000k
'-0)
0C0,
+I+I +I +I +I +I +I+
+I +1 +1+++I
,-
C
Ln L
O mnC
CD
00C
OM
OOn
ZLc
Z
M Ln
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
+1
oo
NNM
N
00
CD
u-,
c) C,c CDlM~
N COt 000
00 N Cl CO N in
0
CO COCO
0)0
COCO
COCO Cl
CO
CO?m
Cl
N
Cl
0i
-0'-'-
o
-n0)
C~O~
000)
0).
.00
rCO.-
0)
0)
0)
0)
0)
0)
a)
V)
4-
CO~n
(1)
(J03
to Q
00
V)
in
ON
N
ON
ON >
vO
N
O
ON
>>
>
N
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
N
>
O
H
44 NORTHWESTERN NATURALIST
surements was calculated. These were then
compared using ANOVAs and t-tests.
We were interested in determiningif morphologically dissimilar murrelets(which could
representdifferentage classes) were caught at
differenttimes of day or with the different
capture methods at Desolation Sound. As a malebiased sex ratio has been found in birds captured using mist-netting,but not night-lighting at this site (Vanderkistand others1999), we
wanted to assess whether this patternwas reflected in other characters. Comparisons
among sites and methods were made using
data fromonly 1997, as this was the only year
work was conducted at Mussel Inlet and when
night-lightingwas used. There were insufficientsamples to examine methods and capture
sessions during any year otherthan 1997. Comparisons among methods and capture sessions
were made using data collected by 1 observer
to avoid inter-observereffectsin this test.
DiscriminantFunctionAnalysis
Discriminant function analysis (DFA) was
conducted using all variables to determine if
male and female marbled murreletscould be
accuratelyassigned to sex using morphometric
characters.The accuracy of the DFA was crosschecked using a jacknife analysis (Tabachnik
and Fiddell 1989). Each variable was then
dropped and the DFA re-runto determineifaccuracy could be improved. A discriminantformula was derived using variables thatresulted
in the most accurate DFA. Due to significant
differencesamong years and sites the DFA was
conducted on data fromonly 1 year and 1 site
(1997 birds at Desolation Sound, as thisdata set
had the largest sample size). Due to significant
inter-observerdifferencesa 2nd DFA was performedon murreletsmeasured by only 1 person (from 1997, the observer with the largest
sample size). DFAs were conducted on adults
and juveniles separately,and between adults
and juveniles within each sex.
82(2)
RESULTS
A total of 937 marbled murreletswere captured from1994 to 1997 at Desolation Sound (n
= 818), and Mussel Inlet (n = 119) during 1997.
Excluding birds that could not be sexed, or
were recaptured (assessed separately), the
morphometricsof 664 murrelets(391 males and
273 females) were available foranalysis (Table
1).
Inter-observer
Effects
Significant differenceswere found among
observers' measurements in some variables
(Fig. 1). Bill width,wing chordlengthand culmen length were the most inconsistentamong
observers (Table 2). Inter-annualcomparisons
of measurements made by the same observer
(observers 1 and 2) were more consistentthan
those found in the entire data set. No significant differenceswere found between years in
any variables measured by observer 1, while
male wing chord lengthand femalebill height
measured by observer2 were significantlydifferentbetween years (Table 2).
The deviation of male wing chord lengthand
female bill width from the mean of all measurementsremained significantlydifferentbetween Mussel Inlet and Desolation Sound
(wing chord length:F3,157= 4.2, P < 0.007; female bill width: F3,98 = 8.7, P < 0.0001; other
variables: P > 0.05).
Sixty-sixmarbled murreletswere recaptured
in differentyears (Table 3). Significantdifferences were found between the 2 measurement
periods in bill height(U = 2983.5,P < 0.05), bill
width (U = 2949.5, P < 0.05), and tarsus length
(U = 411.0, P < 0.05), but not in wing chord
length (t97 = 1.0, n.s.) or culmen length (t96 =
0.06, n.s.). In all cases the morphometricsof
murrelets were smaller upon recapture than
during the original measurement (bill height:
year 1, 6.1 ? 0.4 mm, n = 51; year 2, 5.8 1 0.4
mm, n = 51; bill width: year 1, 4.5 1 0.8 mm,
n = 51; year 2, 4.1 J 0.6 mm, n = 51; tarsus
FIGURE1. Boxplotsofculmenlength,
billheight,
billwidth,wingchordlengthandtarsuslengthmeasured
observersfrom1994through1997.The centerline is themedian,thelengthofthebox is the
by different
range within which the central50% of values fall. The long lines representthe range in which 75% of values
fall. An asterisk denotes an outlier (1.5 times the H-spread or inter-quartilerange) and an open circle represents an extremeoutlier (3 times the H-spread).
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
HULL AND OTHERS: MARBLED
2001
AUTUMN
MURRELET
45
MORPHOMETRICS
8-
2120-
7-
E186
E
119E
S17
5-
16
a" 16
0_E15-
4-
14133
3
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
TIII
l F
0 1 2 3 4 5 6 7 8 9 1011 12 13 14
Measurer
Measurer
150-
6-
E5-
140-
E54
HU
13130-
20
I I I I I I I 120II
1 2 3 4 5 6 7 8 9 10 11 12 13 14
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Measurer
Measurer
25-
E
E 20-
a)
2 1510-
0 1 2 3 4 5 6 7 8 9 1011 121314
Measurer
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
46
NORTHWESTERN
NATURALIST
82(2)
TABLE 2. Statisticalsignificanceofvariationin the measurementsof adult marbled murreletmorphometrics
by differentobservers and by the same observerin differentyears. Level of significanceare indicated by * (P
< 0.05), ** (P < 0.01), ***(P < 0.001); otherwise P > 0.05.
Among observers (number of observers)
Morphometric
Wing chord length:
1995 (n = 3)
1996 (n = 5)
1997 (n = 8)
Withinobservers,
between years
1995 to 1996
1996 to 1997
Males
Females
F2, 70 = 0.36
F2, 36 = 4.0*
F2, 46 = 3.19*
F2, 21 = 0.8
F2, 120
1.1
F6, 68 = 1.02
t17 = 0.6
t14 = 1.1
Males
F2, 70 = 1.64
F2, 46 = 2.03
F2, 120 = 0.6
t17 = 1.0
t28= 1.3
t17= 0.2
H2 = 6.65*
H2 = 13.81***
H4 = 6.13
H2 = 1.19
H7 = 39.5***
F6 = 17.1**
U = 990.0
U = 345.0
U = 487.0
U = 402.5
H2 = 1.87
H2 = 3.48
H4 = 7.57
H2 = 4.4
H7 = 7.8
H7 = 3.09
U = 987.0
U = 328.0
U = 566.0
U = 340.5
H4 = 6.82
H2 = 0.26
H7 = 4.2
H7 = 13.5*"
Culmen length:
F2,37 = 0.44*
Females
Bill width:
Males
Females
Bill height:
Males
Females
Tarsus length:
Males
Females
length: year 1, 19.8 + 2.4 mm, n = 18; year 2,
17.5 ? 1.1 mm, n = 18). The % ME was highest
for wing chord length (66.8), followed by bill
height (55.5), culmen length (53.0), bill width
(44.2), and tarsus length (36.8).
MorphometricComparisons
Adult marbled murrelets were significantly
larger than recently fledged juveniles with 1
exception. Female adults and juveniles did not
differ significantly in tarsus length (Tables 1
and 4).
When deviations from the mean were examined, there also were no significant differences in any morphometric variable among
years in either sex (wing chord length: males
F2, 158 = 2.8, females F2, 99 = 0.2; culmen
length: males F2, 159 = 0.8, females F2, 100 - 0.6;
bill width: males F2, 156 = 0.5, females F2, 99 2.8; bill height: males F2, 156 = 1.3, females F2,
96 = 1.3; tarsus length: males t53 = 1.4, females
t24 = 0.8). Using data from 1 observer revealed
TABLE 3. Recaptures of marbled murrelets from
1994 through1997.
Year
1994
1995
1996
1997
Total
Same year
recapture
2
10
3
26
41
Inter-year
recapture
1
9
10
46
66
F6,68 = 2.4**
F2,21 = 5.4*
Total
3
19
13
72
117
t46 = 2.2*
t14 = 0.7
t44 = 0.02
U = 339.0
U = 391.0
no significant differences in males among capture sessions in any variable (insufficient data
to test in females), or in any variable in both
sexes using different capture methods (Table
5).
Comparison of deviations of birds from Mussel Inlet to those at Desolation Sound indicated
that male murrelets from Mussel Inlet had significantly smaller culmen lengths (t95 = 5.4, P
< 0.0001) and tarsus lengths (t75 = 3.3, P <
0.002) but were not significantly differentin the
other variables (wing chord length t104 = 0.8,
bill width t92 = 0.4, bill height t90 = 0.8). Female murrelets differed significantly between
the 2 sites in culmen length (t112 = 3.0, P <
0.003), wing chord length (tl26 = 2.0, P < 0.05),
and tarsus length (t114 = 3.4, P < 0.001), but not
in bill width (t117 = 1.9, P > 0.05) or bill height
(t117 1.5, P> 0.05).
Discriminant Function Analysis
Male and female murrelets differed significantly in size during all years of the study
(overall, female measurements as a percentage
of male measurements were: wing chord length
98.6%, culmen length 97.8%, bill width 97.8%,
bill height 98.1%, and tarsus length 98.2%; Tables 1 and 6). However, the variables that differed were not consistent each year. During
1994 adults differed only in bill height, during
1995 wing chord and culmen length were different, no variables differed in 1996, and in
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
HULL AND OTHERS: MARBLED MURRELET MORPHOMETRICS
AUTUMN 2001
47
TABLE 4. Statisticalsignificanceof differencesin morphometricvariables between adult and juvenile male
and female marbled murreletsfromDesolation Sound in 1997.
Adults versus juveniles
Males
Females
t155= 7.3, P < 0.001
t155 9.2, P < 0.001
U = 11756.5, P < 0.05
U = 12056.5, P < 0.05
U = 11582.0, P < 0.05
tl09 = 6.5, P < 0.0001
t120= 6.4, P < 0.0001
U = 6971.0, P < 0.05
U = 7173.5, P < 0.05
U = 6778.0, P > 0.05
Variable
Wing chord length
Culmen length
Bill width
Bill height
Tarsus length
1997 wing chord length, bill height, and tarsus
length were significantly different (Tables 1
and 6). Significant differences in bill width, bill
height, and tarsus length were found between
the sexes in juveniles, with males larger in all
variables than females (Tables 1 and 6).
The DFA conducted on 1997 adult murrelets
from Desolation Sound accurately assigned
66% overall to sex (62.0% males and 70.1% females). With cross-validation the accuracy was
reduced to 65% overall (62.0% males and 67.9%
females). The most accurate DFA was achieved
by using all variables. The derived discriminant formula was:
D = 31.5 - 0.5 (culmen length)
- 0.6 (bill height) + 0.6 (bill width)
- 0.2 (wing cord length)
- 0.2 (tarsus length)
When data are applied to this formula, those
>0 are males and those <0 are females.
Using data from only 1 observer improved
the accuracy of the DFA to 76% overall (71.4%
females and 79.4% males), and with cross-validation accuracy improved to 69% (71.4% females and 67.6% males). Leaving out bill width
provided the most accurate result, with the following formula being generated:
D = 53.8 - 0.6 (culmen length) - bill height
- 0.3 (wing)
DISCUSSION
The data collected during this study have
provided the 1st morphometric description of
marbled murrelets from Desolation Sound and
Mussel Inlet. They also have allowed the 1st detailed description of sexual differences in marbled murrelet morphometrics using live birds,
and have allowed assessment of the validity of
a DFA to determine the sex of marbled murrelets.
TABLE 5. Comparisons of marbled murreletmorphometricsmeasured in birds captured using different
methods and during differentsessions (data from1 observer during 1997; data were insufficientto test females among sessions). There were no statisticallysignificantdifferencesin any of the variables tested.
Comparison between
methods
(n: mist-net,night-lighting)
Comparison among
sessions
(n: am, pm, all night)
Males
t23 = 1.2 (22, 13)
F2, 32 = 1.9 (10, 12, 13)
Males
t29 = 0.7 (22, 13)
f2, 32 = 0.9 (10, 12, 13)
t32 = 2.1 (22, 13)
F2, 32 = 2.2 (10, 12, 13)
t32 = 1.5 (22, 13)
F2, 32 = 1.7 (10, 12, 13)
t21 = 0.4 (22, 13)
F2, 32 = 1.1 (10, 12, 13)
Variable
Wing chord length:
Females
Culmen length
Females
Bill width
Males
Females
Bill height
Males
Females
Tarsus length
Males
Females
t4 = 2.2 (18, 4)
t4 = 0.6 (18, 4)
t4 = 1.6 (17, 4)
t3 =1.1(18,
4)
t21 = 0.5 (18, 4)
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
48 NORTHWESTERN NATURALIST
82(2)
TABLE 6. Significanceof variation of morphometricsbetween male and female marbled murrelets.
Between sexes
Variable
Wing chord length
Culmen length
Bill width
Bill height
Tarsus length
Juveniles
1994
1995
1996
1997
t25= 0.8
P > 0.05
t25= 0.2
P > 0.05
U = 151.0
P < 0.05
U - 166.5
P < 0.05
U = 130.5
P < 0.05
t65 = 1.9
P > 0.05
t64= 1.6
P > 0.05
U = 592.5
P > 0.05
U = 445.0
P < 0.05
U - 523.5
P > 0.05
t104= 3.3
P < 0.002
tl07 = 3.1
P < 0.003
U = 1970
P > 0.05
U = 1627.0
P > 0.05
t92= 1.4
P > 0.05
t92= 1.9
P > 0.05
U = 1551.0
P > 0.05
U = 1609.5
P > 0.05
U 1047.5
P > 0.05
t302= 5.5
P < 0.0001
t302= 5.5
P < 0.0001
U = 30790.5
P > 0.05
U = 28939.5
P < 0.05
U = 29213.5
P < 0.05
Inter-observer
Effects
Like many large, long-termprograms, the
study at Desolation Sound requires a number
of observers.However,it has resulted in inconsistenciesin measurements,even thoughall observers in the program were trained in measurement techniques. The varietyof tests carried out in this study indicated the degree of
inter-observereffects.Significantdifferencesin
data fromdifferentobserverswithinyears, the
inconsistent differencesbetween sexes each
year, inter-annualcomparisons from2 observers, and recapture data that suggested birds
had decreased in size, all highlightedthemagnitude of the effects.
Percent ME revealed that the least reliable
measurement was wing chord length and the
most reliable was tarsus length. However, all
morphometricshad a high %ME (>10%, as
suggested by Lougheed and others1991). Studies assessing thereliabilityofmorphometricsin
passerines foundwing lengthto be the mostreliable (Gosler and others1998), which suggests
wings of marbled murreletsmay be more difficultto measure accuratelythan those of many
passerines.
Althoughusing data fromonly 1 observerreduced variabilityand improved the accuracy of
the DFA, it is not feasible foronly 1 observerto
measure all birds in the Desolation Sound program. The importanceofcarefullytrainingpersonnel, cross-checkingdata among observers,
and assessing the degree of inter-observereffectswhen drawing conclusions about significant differencesin morphometrics,cannot be
overstated(as suggested by Barrettand others
1989; Dzubin and Cooch 1992).
The standard deviations in measurementsof
-
marbled murreletsin our study were smaller
than those in a study by Pitoccelli and others
(1995) but were larger than those in a study by
Sealy (1975). This suggests that the use of 1
measurer results in smaller errors; however,
Sealy's (1975) study was conducted in a museum on dead specimens, which may have also
allowed more accurate measurements.
Morphometric
Comparisons
Adult marbled murreletswere significantly
larger than newly fledged juveniles. No differences in morphometricswere found among
birds across years, captured at differenttimes
of day, or captured using differenttechniques.
Therefore, morphologically distinct groups
were not apparent in this study,and no further
insightcan be provided intowhy a male-biased
sex ratiois foundin birds captured by mist-netting but not night-lighting.
Significant differences in culmen length,
wing chord length, and tarsus length were
found between birds at Desolation Sound and
Mussel Inlet, suggesting morphological differences between the populations. Comparisons
of murreletsin the currentstudy to those from
Langara Island (approximately 600 km N of
Desolation Sound) and Alaska (Sealy 1975; Pitocchelli and others 1995) indicated theywere
similar in size, with some exceptions (Table 7).
Birds in the currentstudy had larger culmen
lengths and bill heights,but smaller tarsi than
those at Langara Island. They also had longer
wing chords and culmens than those in Alaska.
However, comparisons to the Alaska data set
are difficultbecause the sexes were combined
in that study and wing chord length differed
between ground- and tree-nestingmurrelets.
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
HULL AND OTHERS:MARBLEDMURRELETMORPHOMETRICS
AUTUMN2001
ft
The distribution of marbled murrelets in
North America is described as discontinuous
with possibly geneticallydiscrete populations
and with unknown rates of exchange among
them (Ralph and others 1995). While the morphometric data that are provided give an assessment of the discreteness of populations of
marbled murrelets,statisticalcomparisons are
not possible until the degree of inter-observer
differencesamong the studies is determined.
Until such effectsare determined and minimized, conclusions about differencesamong
these populations must be tentative.
B1
LO
>1
N
o
0
+1+1
a
cn
cnC)
?d
a,
bo
N
N
n
UN
UN
l
N
a,
LN
00t
oO
CC
cc)
d
B
4-1
a
a
Nl
.-S
a
C)e
+1
+
+
C)l
C)l
CD)
LCL
+C -+DL 110- 00
+i 0
+1
C)
SB
Sa4-a
l
a)
Y
Cr
00'
r)
CIA
0
)
B )
e
,
0
c
a
CC
otia
.a
kQ)
C)
C)
Sd
0
B0
DiscriminantFunctionAnalysis
The sexual dimorphismfound in adult marbled murrelets was small (females were approximately98% thesize ofmales), but allowed
a 65 to 66% success rate in separating sexes using a DFA, with the sex of adults being more
accurately identifiedthan the sex of juveniles.
Generally the sex of alcids is very difficultto
determine in the field due to monomorphic
plumage and similarityin body size (Bedard
1985). Reliable determinationof the sex of alcids using morphometricshas been achieved
with crested auklets (Aethiacristatella;95% accuracy using bill shape, Jones 1993), Cassin's
auklets (Ptychoramphusaleuticus; 94% and 86%
0,1
'Z
49
?i+ +I
+I ? ?
+I
+1
+1a'z++1
'N+a
n I
Cc7
a
accuracy,Nelson 1981 and H. Knechtelunpubl.
data, respectively),ancient murrelets(Synthliboramphusantiquus; using bill depth 87% and
70% accuracy,Jones1985 and Gaston 1992, respectively),and Atlanticpuffins(Fraterculaarctica; 65% accuracy using bill depth and culmen
length,Corkhill 1972). There is littlesexual dimorphismin common murres(Uria aalge) making the use of a DFA impossible (Threlfalland
Mahoney 1980), and therealso is littleapparent
dimorphism in razorbills (Alca torda),black
guillemots (Cepphusgrylle),and least auklets
(Aethiapusella) (Gaston and Jones1998). While
the accuracy of the DFA in marbled murrelets
is greaterthan in some otheralcids, it has only
limited value (particularly given the problem
of inter-observereffects),and in order to eliminate errors in the assignment of sex, the molecular technique (Vanderkistand others 1999)
must be employed.
This study has determinedthattheremay be
some morphological differencesbetween birds
fromMussel Inlet and Desolation Sound, and
apparent differencesbetween the birds in our
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
50
NORTHWESTERN
NATURALIST
82(2)
study and those in populations in Langara Island and Alaska. The small degree of sexual dimorphism in the species prevents the widespread use of a discriminant formula to identify the sex of individual birds. Future studies
of marbled murrelets need to minimize interobserver effects by very careful training of personnel and by cross-checking data from all observers.
ACKNOWLEDGMENTS
We thank the numerous field assistants who have
worked on this program. T. D. Williams, S. Courtney
and C. Thompson provided comments that improved earlier drafts of the manuscript. Forest Renewal B.C. (FRBC), NSERC, Chair of Wildlife Ecology at Simon Fraser University,Canadian Wildlife
Service,MacMillan-Bloedel Ltd, Timber/WestForest
Ltd, InternationalForest Products Ltd, WesternForest Products Ltd, and Pacific Forest Products Ltd.
gave generous financialsupport. Workwas conducted under permitsBC SCI 97/012A and101201 AQ.
LITERATURE CITED
BAILEYRC, BYRNESJ. 1990. A new, old method for
assessing measurement error in both univariate
and multivariatemorphometricstudies. Systematic Zoology 39:124-130.
BARRETT
RT, PETERZ
M, FURNESS
RW, DURINCK
J.
1989. The variabilityof biometricmeasurements.
Ringing and Migration 10:13-16.
G. 1992. Systematics,biodiversity
BARROWCLOUGH
and conservationbiology. In: Eldredge N, editor.
Systematics,ecology and the biodiversitycrisis:
New York,NY: Columbia UniversityPress. p 121143.
BEDARDJ.1985. Evolution and characteristicsof the
AtlanticAlcidae. In: NettleshipDN, BirkheadTR,
editors. The Atlantic Alcidae: London, UK: Academic Press. p 1-51.
BURNSRA, KAISERGW, PRETASHLM. 1995. Use of
mist nests to capture marbled murreletsover the
water.NorthwesternNaturalist 76:106-111.
CARTERHR, STEIN JL. 1995. Molts and plumages in
the annual cycle of the marbled murrelet. In:
Ralph CJ,Hunt GL, Raphael MG, PiattJF,editors.
Ecology and conservationof the marbled murrelet: General Technical Report PSW-GTR-152,Albany CA: USDA ForestService,PacificSouthwest
Research Station. p 99-109.
CORKHILLP. 1972. Measurements of puffinsas a criteria of sex and age. Bird Study 19:193-201.
NEE K, JALILAK, KENG WL. 1994.
COUNSILMANJJ,
Discriminant analysis of morphometriccharacters as a means of sexing mynas. Journalof Field
Ornithology65:1-7.
DZUBIN A, COOCH E. 1992. Measurement of geese.
General field methods. Sacramento, CA: California WaterfowlAssociation. 20 p.
GASTONAJ.1992. The ancientmurrelet.London, UK:
Poyser.249 p.
GASTON AJ, JONES IL. 1998. The auks. Oxford, UK:
Oxford UniversityPress. 349p.
GOSLER AG, GREENWOOD JJD,BAKER JK, DAVIDSON
NC. 1998. The field determinationof body size
and condition in passerines: a reportto the British Ringing Committee.Bird Study 45:92-103.
HAMER KC, FURNESSRW. 1991. Sexing great skuas
Catharactaskuaby discriminantanalysis using external measurements.Ringing and Migration 12:
16-22.
HEDD A, GALES R, BROTHERS N. 1998. Reliability of
morphometricmeasures fordeterminingthe sex
of adult and fledgling shy albatrosses, Diomedea
cauta cauta,in Australia. WildlifeResearch 25:6979.
JONESIL. 1985. Structureand functionof vocalizations a related behavior of the Ancient murrelet
antiquus) [thesis]. Toronto,ON:
(Synthliboramphus
Universityof Toronto.178 p.
JONESIL. 1993. Sexual differencesin bill shape and
external measurements of crested auklets Aethia
cristatella.Wilson Bulletin 105:525-529.
KAISER GW, DEROCHER AE, CRAWFORD S, GILL MJ,
MANLEYIA. 1995. A capture technique for marbled murreletsin coastal inlets. Journalof Field
Ornithology66(3):321-333.
KOFFIJBERGK, VAN ERDEN MR. 1995. Sexual dimor-
carbosinenphism in the cormorantPhlalocrocorax
sis: possible implicationsfordifferencesin structural size. Ardea 83:37-46.
LOUGHEED SC, ARNOLD TW, BAILEY RC. 1991. Mea-
surementerrorof external and skeletal variables
in birds and its effecton principal components.
Auk 108:432-436.
NELSON KS. 1981. Sexual measurements of Cassin's
auklet. Journalof Field Ornithology52:233-234.
PHILLIPS RA, FURNESS RW. 1997. Predicting the sex
of parasitic jaegers by discriminantanalysis. Colonial Waterbirds20:14-23.
PITOCCHELLI J,PIATT J,CRONIN MA. 1995. Morpho-
logical and genetic divergence among Alaskan
murrelets.Wilson
populations of Brachyramphus
Bulletin 107:235-250.
PRATER AJ,MARCHANT JH, VUORINEN J.1977. Guide
to the identificationand ageing of Holarctic waders. London, UK: British Trust for Ornithology.
168 p.
RALPH CJ, HUNT GL, RAPHAEL G, PIATT JF. 1995.
Ecology and conservationof the marbled murrelet in North America: an overview. In Ralph CJ,
Hunt GL, Raphael MG Piatt JF,editors. Ecology
and conservation of the marbled murrelet.General Technical Report PSW-GTR-152.Albany,CA:
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions
AUTUMN
2001
HULL AND OTHERS: MARBLED
USDA Forest Service, PacificSouthwest Research
Station. p 3-22.
SEALY SG. 1975. Aspects of the breeding biology of
the marbled murreletin British Columbia. Bird
Banding 46:141-154.
TABACHNIKBG, FIDDELL LS. 1989. Using multivariate statistics.New York:Harper Collins. 746 p.
THRELFALL W, MAHONEY SP. 1980. The use of mea-
surements in sexing common murres fromNewfoundland. Wilson Bulletin 92:266-268.
VANDERKIST BA., XUE X-H, GRIFFITHS R, MARTIN K,
MURRELET
MORPHOMETRICS
51
BEAUCHAMPW, WILLIAMSTD. 1999. Evidence of
male-bias in capture samples of marbled murrelets from genetic studies in British Columbia.
Condor 101:398-402.
WHITWORTH DL, TAKEKAWA JY, CARTER H, MCIVER
RW. 1997. A night-lightingtechnique for at-sea
capture of Xantus' murrelets. Colonial Waterbirds 20:525-531.
Submitted24 June2000, accepted24 January
2001.
Editor:C. J.Ralph.
Corresponding
This content downloaded from 142.58.26.133 on Wed, 27 May 2015 17:51:02 UTC
All use subject to JSTOR Terms and Conditions