Content uploaded by Steve Redpath
Author content
All content in this area was uploaded by Steve Redpath on Dec 15, 2017
Content may be subject to copyright.
1RlS
1.39:
152-158
Variation in the male territorial hoot
of
the Tawny Owl
Strix
aluco
in three English populations
B. M. APPLEBY’
&
S.
M.
REDPATH*
Wildlife Conservation Research Unit, Department of
Zoology,
Oxford University,
South Parks Road, Oxford
OX1
3PS,
UK
lnstitute
of’
Terrestrial Ecology, Monks Wood, Abbots Ripton, Cambridgeshire
PEI
7
2LS, UK
Little is known about owl song. We made sonagrams of the territorial calls of 50 male
Tawny Owls
Strix
aluco
from three different areas.
Six
temporal and
four
frequency mea-
sures of the calls were recorded from the sonagrams. The measures of the calls were then
subjected to analysis to try to separate between individual owls and between owls from
different areas. We also looked for similarities between calls of neighbouring owls and for
any effect of habitat on owl hoots. Individual owls were separated on the basis of their
hoots with a high degree of success
(98.6%
overall), and there were significant differences
between areas. Differences were found between calls in woodland and farmland habitats,
but these differences were not in the direction expected to increase sound transmission.
Calls of neighbouring owls did not resemble each other more than calls from owls that
were not in vocal contact, implying that
if
calls are learned by Tawny Owls, they are
learned before dispersal.
When a bird sings. the two fundamental questions asked by
behavioural researchers are
(1)
what advantage does the
bird
gain from singing? and (2) why should other birds re-
spond to the signal in the way they do?
In
order to address
these questions, it is necessary to know what information
the song contains. In order to know this, it is useful to know
whether the song was learned before
or
after dispersal. Song
that developed before dispersal might be similar in closely
related birds, giving information that may be used
for
kin
recognition and mate choice. Song that developed after dis-
persal might result in local dialects and adaptation of song
to habitat.
Much has been written about song development. dialects
and environmental influences
on
song structure in passer-
ines (e.g. Thorpe 1958, Marler 1970, Nottebohm 1975,
Baptista
&
Johnson 1982, Mundinger 1982, Anderson
&
Connor 1985. McGregor
&
Thompson 1988, DeWolfe
et
al.
1989, Williams
&
Slater 1993). Few studies have addressed
these issues in nonpasserines, possibly because song learn-
ing is
not
thought to be common in nonpasserine species.
Song learning has been documented in only four families of
nonpasserines compared with
3
3
families of passerines
(Kroodsma 1982), and some nonpasserine species have been
shown not
to
need learning to develop normal song (dove
species: Lade
&
Thorpe 1964, Nottebohm
&
Nottebohm
1971:
domestic fowl: Konishi 1963). However, song learn-
ing and dialects have been shown to occur in the Bobwhite
Quail
Colinus
virginiaus
(Goldstein 1978. Bailey
&
Baker
1982). Another reason that nonpasserine song has received
so
little attention is that the emphasis in the past
has
been
towards qualitative description of syllable types, whereas the
more simple song of nonpasserines requires quantitative
analysis by sonagrams to investigate individual and popu-
lation characteristics.
In
this paper, we present the results of a quantitative
study of the male territorial hoot of the Tawny Owl
Strix
aluco.
Galeotti and Pavan (1991) showed that individual
male
Tawny Owls could be distinguished on the basis
of
temporal characteristics
of
this hoot, and Galeotti and
Pa-
van (1993a) claimed that males respond more strongly to
playback of territorial hoots
of
strangers than to hoots of
neighbours. This implies that Tawny Owls make use of the
information in the hoot to distinguish individuals.
We made recordings from Tawny
Owls
in three areas of
southern England. Two of the areas were woodland habitat
and one was farmland habitat. We made sonagrams of the
hoots and took six temporal measures and four frequency
measures from each one. We wanted
to
examine whether
there were differences between the three areas and,
if
so,
whether hoots in woodland were different than hoots in
farmland. The acoustic adaptation hypothesis (Williams
&
Slater 1993) proposes that bird songs that are transmitted
more effectively through a particular habitat are more likely
to be heard and learned by other birds in that habitat, lead-
ing to song becoming adapted to the environment in which
a bird lives. Adaptation of Tawny Owl song to the habitat
could therefore be taken to be indirect evidence of post-
dispersal song learning. The acoustic adaptation hypothesis
would predict that owls in dense woodland should have calls
at a lower frequency than those in open farmland because
152
1Y97
VARIATION
IN
TAWNY
OWL
HOOTS
153
lower frequencies travel better through obstacles such
as
leaves and branches. If post-dispersal learning is occurring
in owl populations, then we would also expect the hoots of
neighbouring birds to be more similar than the hoots of
birds not in vocal contact.
METHODS
Studv sites
Our three recording sites were in southern England at Wy-
tham
Woods.
Oxfordshire (S1"46'N, l"2'W); Monks Wood,
Cambridgeshire (S2"24'N,
O"14'W)
and the Fens,
Cambridgeshire (52"29'N,
0"l'W).
Wytham Woods is an
area of
5
2
5
ha of mixed deciduous woodland surrounded
by
farmland. Monks Wood is
an
area of
148
ha of similar
woodland, and the Fens study area is 16,000 ha of open
farmland, with woods covering <0.5% of the area.
Average Tawny Owl territory sizes were
22
ha in Wytham
Woods
(B.
Appleby, unpubl. data) and
15
ha in Monks
Wood.
Home range size in the Fens
was
126 ha (S.M. Red-
path. unpubl. data). The centre of the Monks Wood study
site
was
about 20 Irm from the centre of the Fens study site.
The centre of Wytham Woods was over
150
km from both
the other sites.
Recordings were made of the hoots of
23
male owls from
Wytham Woods,
10
from Monks Wood and
17
from the
Fens. In all three sites, these accounted for nearly
all
the
owls present in the study area.
Recording equipment and methods
Recordings were made using
a
Uher or Sony Walkman Pro-
fessional tape recorder
(TC-D5
PRO) with
a
Sennheiser
MZW
8
16
microphone. Recordings were made on calm, dry nights
from October to December 1992 and 1993 in Wytham
Woods and in March 1993 in Monks Wood and the Fens.
Birds were stimulated
to
hoot using playback of
an
unfa-
miliar
male
owl. Recordings were made
as
near to birds
as
possible, at
5-50
m.
Sonagram production
Sonagrams were made on
a
Macintosh LCII computer. Son-
agrams
of
the whole hoot were made using Soundedit Pro
software (Macromind Paracomp, Inc.,
San
Francisco, Cali-
fornia. USA), and temporal measures were read off the
screen. Soundedit Pro did not give accurate frequency mea-
sures,
so
sonagrams
of
the first note of each hoot were
made with Canary software (Cornell Laboratory of Orni-
thology, Ithaca. New York, USA), and the frequency mea-
sures were read off the screen.
The temporal measures used were similar to those de-
scribed by Galeotti and Pavan (1991). Six temporal mea-
sures were recorded for each call (Fig.
l):
note
l
(Dl),
in-
ternote interval one
(I4),
note two (D2). internote interval
tno
(15)
and note
3,
which was split up into frequency mod-
Figure
1.
measures. Sonagram
of
Tawny
Otd
hoot
showing
time
and
lrequency
ulated length (FML) and tail. Frequency measures used were
the highest and two lowest frequencies
of
the first note
(HIGH, LOW1 and LOWZ) and the middle of the highest
part of the first note (MED). All time measures were record-
ed in milliseconds, and all frequency measures were record-
ed in kilohertz.
Analysis
Measures were taken from two to five hoots per owl (mean
=
4).
For all analyses, except the discrimination of individ-
ual
owls,
a
mean was taken of the measures for each owl
to
give an average hoot for each owl.
Owls in the Fens were classed
as
out
of vocal contact
when their territories were over
5
km apart with
at
least
two intervening territories between them. Owls in Wytham
Woods were considered out of vocal contact if there were
more than 2
Itm
and three intervening territories between
them and their territories were in different sections
of
the
wood. Wytham Hill separates the three main sections
of
the
wood, providing an acoustic barrier. Monks Wood is much
smaller than the other two study sites, so we had to class
owls
as
out of vocal contact if there were more than
1
Itm
and
two intervening territories between owls. Tawny Owls
can hear and answer each other at
a
distance
of
1.5
km
over open land (Anderson 1961). Although sound does not
travel
as
far in woodland, there was
a
slight chance that
owls classed as out of vocal contact in Monks Wood might
be able to hear each other occasionally. Because the three
sites were analysed separately, any vocal contact between
Monks Wood owls would not affect the results for the other
two sites.
STASTISTICA software (Statsoft, Inc., Tulsa, Oklahoma,
USA) was used for discriminant analysis of individuals and
populations and for
ANOVA
tests.
SAS
software (SAS Institute
Inc., Cary, North Carolina, USA) was used for principal com-
ponent analysis (PCA) to look for similarities between neigh-
bours.
154
B. M.
AI’PI,ERY
I
S.
M.
KEDPATH
IBIS
139
Table
1.
Owls for pach of the ten measures of the
hoot
Results of univariate
ANOVA
between individual Tawny
P-value
Table
3.
tions
separating individual Tawny Owls
Canonical coefficients of the
first
two discriminant func-
Root
1
Root
2
I5
I4
TAIL
D1
D2
FML
LOW
1
HIGH
LOW2
MEU
27.3
26.4
14.5
18.2
5.1
21.1
7.7
36.2
10.4
44.3
<0.001
<0.001
<0.001
<0.001
<0.001
10.001
<0.00
1
<0.001
<0.001
<0.001
MED
I4
I5
FML
D1
TAIL
LOW2
D2
HIGH
LOW
1
-0.52
-0.19
0.63
0.07
0.18
0.31
0.23
0.27
-0.39
-0.06
-0.23
0.71
-0.61
-0.05
0.08
-0.03
0.25
-0.37
-0.42
0.14
RESULTS
Discrimination
of
different areas
Discrimination
of
individual
owls
Analysis of variance showed that all the parameters differed
significantly between individual
owls
(Table
1).
Discriminant
analysis on all ten measures allowed
98.6%
of hoots to be
attributed to the correct individual. There was some redun-
dancy in parameter selection because several parameters
were significantly intercorrelated (Table 2).
All
nine discrim-
inant functions generated were statistically significant
(P
<
0.01).
MED
and
I5
contributed most to the first discriminant
function, measures I4 and I5 to the second (Table
3).
The
first
canonical axis was plotted against the second and
shows a clear separation between owls (Fig.
2).
When the
three areas were considered separately, different measures
were important in discriminating individuals in each,
so
no
measure seemed to be consistently the most important in
discriminating individuals.
Univariate analysis of variance showed
a
significant differ-
ence between the areas only for the frequency measures
LOW1, LOW2 and
MED
(Table
4).
MANOVA
between areas
using all ten measures of the hoots showed the areas were
significantly different (Wilks lambda
2.46,
P
<
0.001). Dis-
criminant analysis using five
of
the ten measures allowed
owls to be classified into the correct area with
8 1.6%
suc-
cess. Both discriminant functions were statistically signifi-
cant
(P
<
0.02), with the first function accounting for most
of the variance. LOW2,
I5
and LOWl contributed most to
the first discriminant function, and MED and
I4
contributed
most to the second (Table
5).
Figure
3
shows owls from the
three areas plotted on a graph of the first canonical axis
against the second canonical axis. The first axis weakly sep-
arates Wytham Woods from the other two sites, and the
second axis separates the Fens from Monks Wood.
Table
2.
flip
two rnensures
Cp
<
0.05)
Correlations between the different measures of the Tawny Owl calls. Underlined digits indicate a significant correlation between
I5
I4
TAIL
D1
D2
FML
LOWl
HIGH
LOW2
MED
I5
I4
TAIL
D1
1)
2
FML
LOW
1
HIGH
LOW2
MED
1
0.26
0.26
0.19
-
-
-
-0.24
-0.2
-0.28
-0.23
-0.25
-0.28
-
-
-
-
-
1
-0.22
-0.13
-0.17
0.07
0.12
-0.02
-
-
0.2
0
-
1
0.5
0.12
-0.21
-0.06
-0.19
-0.03
-0.18
-
-
-
1
0.15
1
0.16 0.1
1
-
-0.21
-0.16
-0.15
1
-0.22
-0.09
0.09
0.16
1
-0.17
-
-0.19
-0.14
-
-1
-0.22
-0.1
0.05
0.4
0.9
-
-
~
-
0.67
0.5
1
-
-
0.31
1997
VARIATION
IN
TAWNY OWL
HOOTS
155
15
10-
5-
4
”-
-5
-
-10
I
+
+
6
8;
,
Habitat differences
Multivariate pairwise
MANOVAS
between the three areas sep-
arately showed that, after Bonferroni adjustment for multi-
ple testing, the Fens was significantly different from Wytham
and Monks Woods but Wytham and Monks Woods were not
significantly different from each other.
Because the Fens was of a different habitat type to Wy-
tham and Monks Woods,
a priori
contrasts were included in
univariate tests to compare the Fens (farmland site) with
the two woodland sites. Four of the ten measures,
FML,
LOW1.
HIGH
and LOW2, were significantly different be-
tween the Fens and the woodland sites (Table
6).
The fre-
quency measures that showed a significant difference be-
tween woodland and farmland sites were not consistently
lower in the woodland sites. HIGH conformed to the pattern
expected and was significantly higher in the Fens, but
LOW2
was lowest in the Fens and LOW1 was intermediate
in
the
Fens. This implies the “shape” of the first note varied be-
tween habitat types, being more “curved” in the Fens. The
samc recording equipment was used at each site,
so
this
difference was not an effect of the recording microphone.
2-
1-
N
2
0-
1-
2-
Table
4.
of thc
tm
measures of the Tawny Owl hoot
Rrwlts
of
univariate
ANOVA
between study sites for each
F-ratio,,,, P-value
14
1.10
ns.
I5
1.96
ns.
TAIL
0.72
ILS.
Dl
1.34
ns.
DL
2.88
ns.
FML
3.07
ns.
LOW
1
16.0
<0.001
HIGH
2.97
ns.
r,owL
19.5
<0.001
ME11
4.38
<0.02
Table
5.
separating the three Tawny Owl study sites
Canonical coeJicients of the two discriminant functions
~~ ~~ ~~
Root
1
Root
2
LOW2 0.78
0.41
I5
0.58
0.06
LOW
1
0.53
0.3
3
MED
0.21 0.83
I4
0.06
0.61
Similarities between neighbours
Principal
component
analysis
(PCA)
The owl hoots from each area were plotted in multidimen-
sional space using
PCA
of the temporal measures. More sim-
ilar hoots are found closer together
on
the
PCA
map. The
distance between hoots
on
the
PCA
plot was then calculated
for every pair of owls that had adjoining territories and ev-
ery pair of owls that was not in vocal contact according to
the criteria stated in the Methods. In none of the three study
sites were the plotted distances between hoots of neigh-
bouring pairs significantly different from the plotted dis-
tances between pairs of males that were not in vocal contact
(Table
7).
This implies that the hoots of neighbouring owls
are not more (or less) similar than hoots of owls that do
not have vocal contact.
Dlflerences
bet
ween pairs
For each measure of the hoot, the magnitude of the differ-
ence was calculated between pairs of neighbouring owls and
pairs of males that were not in vocal contact (strangers).
Univariate
ANOVA
was performed to examine
if
the differ-
ences for the stranger pairs were larger. If
so,
it would imply
neighbouring pairs had more similar hoots. There was
no
significant difference between the neighbour and stranger
groups in any of the measures at any of the areas (Table
81.
0
+
#
++
+
++e
+++
+
++
+
0 0
+
%o
000
0
on
3‘
Axis
1
0
Fen
0
Monkswood
+
Wytham
Figure
3.
Average hoot
for
each Tawny Owl plotted
on
tirst canonical
axis against second canonical axis showing separation between the sites.
1
5
6
8.
M.
APPLEBY
YU
S.
M.
RBDPATH
IBIS
139
Table
6.
ti!!
Owl
hoot
between woodland sites and the Fens
Kesults
of univariate
ANOVAS
on
u11
measures
of
the Taw-
~ ~ ~~ ~~ ~~
F-ratio, 4h P-value
I4
0.44
ns.
I5
3.78
n.s.
TAIL
0.81
ns.
D1 0.41
n.s.
D2 0.38
ns.
FML
6.14 <0.02
LOW
1
4.1
5
<0.05
HIGH
4.53 <0.04
LOW 2
12.41 <0.01
ME11
3.06
ns.
Variability
of
hoots
in
the
three populations
The coefficient
of
variation was calculated
for
each measure
in each area because this is a standardized measure of vari-
ability (Sokal
&
Rohlf
1981).
A
Friedman’s test was per-
formed on the ten coefficients of variation between areas to
test whether any of the sites had a higher variability. The
variation was significantly higher at the Fens (Friedman’s
xAz
=
6.2,
P
<
0.04)
(Table
9).
DISC
ITS
S
ION
After we examined six temporal and four frequency mea-
sures of their hoots, we could distinguish Tawny Owls from
each other with
a
high degree of accuracy. There
was
no
indication that any part of the hoot that we measured con-
sistently acted as an individual “identifier” for the owls.
No
measure was consistently important in discriminating indi-
viduals in
all
three areas: for example
I5
was important in
discriminating individuals in Wytham Woods and the Fens
but was not important
in
Monks
Wood.
In
addition, Galeotti
and Pavan
(1991)
described
FML
as being very important
Table
7.
Average distances between hoots on
a
PCA
“map” for
groups
uj
neighbouring Tawny Owls and groups of Tuwng Owl
strririgers
/ram
all
three sites.
A
small
PCA
distance implies the hoots
ari’ similar to each othrr
Mean
distance
Site
Owl group
n
(cm)
s.e.
P-value
~~ ~
Wytham
Woods
Neighbour
32 2.65 0.25
Monks
Wood Ncighbour 14 2.71 0.42
Fen5
Neighbvur
25 3.45 0.52
Stranger
32 2.98 0.23
n.s.
Stranger
9
3.31 0.55
ns.
Stranger
53 3.02 0.27
n.s.
Table
8.
Tables showing
nieuti
time differences it1 hoot incasures
for groups
of
neighbour pairs and groups of slrangcr pairs of
Tciwny
Owls at
the
three sites, and results of an
ANOVA
between
the
two
groups (none of the F-values was significant)
Mean difference
Variable Neighbours Strangcrs
d.E
I:
Wytham
Woods
Dl
116.6
14 662.82
D2 32.04
I5 68.08
FML 126.07
TAIL 169.42
LOW
1
54.8
HIGH
45.3
LOW2
41.3
MED 38.7
Monks Wood
D1
83.8
I4 605.6
D2
28.2
I5
66.8
FML 43
TAIL
116.9
LOW1
41.7
HIGH 70.8
LOW 2 56.3
MED 64.3
Fens
D1 171.9
I4 925.6
D2
21.3
I5
107.5
FML 143.5
TAIL 177.2
LOW
1
59.4
HIGH 81.6
LOW2 52.6
MED 80.6
12
3.02
553.7
41.57
58.63
125.64
142.31
51.8
80.8
54
64.8
76.7
35.8
64.6
35.1
136.7
74
90.5
49.6
94
719
128.8
791.8
22.1
99.3
158.9
143.1
41.6
76.8
5
7.4
69.9
1.63
1.63
1.63
1,63
1.63
1.63
1.24
1,24
1,24
1,24
1.21
1.21
1.21
1,2
1
1,21
1.21
1,9
1,9
1,9
1,9
1,77
1,77
1.77
1,77
1.77
1.77
1,19
1,19
1,19
1,19
0.07
1.16
2.1
0.5
0
0.82
0.04
3.4
0.88
2.76
0.06
0.36
0.38
0.01
0
%5
0.43
1.04
0.28
0.0
3
0.67
3.32
1.03
0.05
0.15
0.19
1.77
0.73
0.03
0.04
0.13
in discriminating individuals in their Italian study sites be-
cause it was highly variable between owls but very constant
in the hoots of each individual owl. Amongst the English
study sites,
FMT,
was important only in discriminating in-
dividuals in
Monks
Wood.
Hoots from the three areas we studied could be distin-
guished with a high degree of accuracy using multivariate
measures. There was some evidence that habitat irilluenced
hoot structure.
Hoots
from the Fens were significantly dif-
ferent from hoots from each of the two woodland sites, and
hoots from the woodland sites were not significantly differ-
ent from each other. In addition, four of the ten measures
differed significantly between the Fens and the woodland
sites. Three of the four measures which were significantly
different between the Fens and woodland sites were fre-
1997
VARIATION IN
TAWNY OWL
BOOTS
157
Table
9.
(it
earh
o/
the
thrre
study
sitps
lor
Tawny
Owl hoots
Mem
o/
thr
cn@cimls
yf
variancefiv
all
ten
measures
Sitc
Coeff. var.
n
Fcns
0.142
10
Monks Wood
0.101
10
Wvtham
Woodq
0.125
10
quency measures. The acoustic adaptation hypothesis would
predict that these frequencies would be lower in woodland
because low frequencies travel better through obstacles such
as branches and leaves. This was the case for only one of
the three frequencies that differed. There might be a con-
founding factor in that the owls in the Fens were much
farther apart than owls in the woodland areas.
Low
fre-
quencies travel farther than high frequencies,
so
they might
also be favoured in the Fens. It is therefore difficult to be
certain that accoustic adaptation was not taking place.
We found no evidence that owls with neighbouring ter-
ritories had more similar hoots. It is possible that if post-
dispersal learning was taking place, owls would learn from
only one neighbour. Because each owl was compared with
each
of
its neighbours, groups of neighbours
in
our analysis
would then contain some similar hoots and some hoots that
were no more similar than those of owls that were not in
vocal contact. Because our sample sizes were large, however,
we would still expect to see reduced differences between
hoots in groups of neighbours
if
learning was taking place.
No
such trend was found,
so
we concluded there was no
evidence for learning from neighbours.
If
owls were learning their hoots after dispersal, we would
expect to find that the three areas were distinguishable and
that there were differences between hoots in different hab-
itats. We would also expect that, on average, neighbours had
more similar hoots than owls with no vocal contact and that
hoots in the Fens were more variable because the owls have
less vocal contact. We did find that the populations were
distinguishable and there were habitat differences, although
these were in the “shape” of the first note rather than sim-
ply the frequency, and we did find hoots in the Fens were
more variable. However, the most direct measure
of
post-
dispersal learning was whether owls with territories close to
each other had more similar hoots. There was no evidence
for this, despite a thorough analysis using two different ap-
proaches.
so
we concluded post-dispersal learning was not
taking place.
If
Tawny Owls do not learn their hoots after dispersal,
they might inherit or learn their hoots from their fathers or
learn their hoots from a neighbouring male before dispersal.
Fledged
owls
remained
on
the natal territory for approxi-
mately
4
months after fledging,
so
they had ample time to
learn their calls. Owl chicks perform squeaky hootlike calls
before leaving their natal territories
(B.M.
Appleby
&
S.M.
Redpath, pers. obs.). If owls acquire their hoots before leav-
ing the natal territory and then disperse short distances be-
fore setting up territories, this behaviour could lead to a
number of owls in an area having similar hoots. Over
83%
of the young Tawny Owls in one British population dis-
persed under
10
km before setting up territories (Cramp
et
al.
1985).
The increased similarity
of
hoots within areas
that we revealed could have arisen in this way The different
shape of the first note in the Fens could also be simply a
result of a local variant of the first note. The increased vari-
ability in farmland could occur if owls acquired their hoots
before dispersal and there was increased dispersal in farm-
land habitat. Unfortunately, no information on dispersal dis-
tances of Tawny Owls in different habitats is available.
Radio-tagging juveniles before dispersal and comparing
sonagrams of their hoots once they have established terri-
tories with the hoots of their fathers and males with neigh-
bouring territories to their fathers would show whether the
young acquire their calls before dispersing, and from where
they acquire them. However, because
of
the high mortality
of young Tawny Owls
(G.
Hirons
1974,
unpublished DPhil
thesis, Oxford University), the amount of time they might
live and the uncertainty
of
the influences they might en-
counter before establishing
a
territory, this field experiment
would be very difficult. Rearing captive owls to investigate
if they will learn song would be preferable, but hand-reared
Tawny Owl chicks do not develop
a
hoot when kept in cap-
tivity
(P.
Galeotti, unpubl.).
We are grateful
to
P.
Johnson for help with statistical analysis and
to
1.
Newton,
R.
Gutierrez.
F!
Galeotti.
D.
Macdonald and
C.
Sillero
Zubiri for commenting on drafts of the manuscript.
B.M.A.
was
financed during this study by a Natural Environment Research
Council studentship.
REFERENCES
Anderson, M.E.
&
Connor.
R.N.
1985. Northern Cardinal song in
three forest habitats in eastern Texas. Wilson Bull. 97:
436-449.
Anderson.
T.
19
61.
En nordsjdansk Natungle-bestand
Str-ir
aluro
i yngletiden
[A
population
of
Tawny Owls in northern Zccland
studied
in
the breeding season]. Dansk Orn. Foren. Tidsskr.
55:
1-55 [in Danish].
Bailey,
K.
1978. ‘The structure and variation
of
the separation call
of the Bobwhite Quail. Anirn. Behav. 26: 296-303.
Baptista,
L.E
&
Johnson,
R.B.
1982. Song variation in insular and
mainland California Brown Creepers
(Certhiii
/lmilfiaris).
J.
Orn.
Cramp,
S..
Brooke, D.J.. Dunn,
E..
Gillmor.
R.,
Hollom.
L?A.L).,
Hud-
son.
R..
Nicholson, E.M., Ogilvie, M.A.. Olney,
P.1.S..
Roselaar,
C.S..
Simrnonds.
K.E.L.,
Voous.
K.H.,
Wallace, D.I.M.. Wattel.
J.
&
Wil-
son.
M.G. (eds). 1985. Handbook
of
Birds
of
Europe. the Middle
East and North Africa. The birds
of
the Western Palearctic.
Vnl.
4.
Oxford: Oxford IJniversity Press.
DeWolfe,
B.,
Baptista,
L.F.
&
Petrinovich.
L.
1989. Song develop-
ment and territory establishment in Nuttali’s White-crowned
Sparrows. Condor
91:
397-407.
Galeotti,
P.
&
Pavari.
G.
1991. Individual recognition
of
male Taw-
123: 131-144.
158
B.
M.
APPLEBY
&
S.
M.
RFIIPATH
IBIS
139
ny
Owls
(Strix
cduro)
using spectrograms
of
their territorial calls.
i;aleotti.
P
&
Pavan.
G.
1993a.
Differential responses
of
male Taw-
ny Owls
Str-ix
n1uc.o
to
the hooting
of
neighbours and strangers.
lhis
135:
300-304.
c;aleotti.
I?
&
Pavan,
G.
1993b. Individually distinct hooting in
mole I'ygmy
owls
Gkucidiurri
pcisscrirltrrrr:
A multivariate
ap-
proach. Ornis Scand.
24:
15-20.
(;oldstein.
1t.K.
1978.
(kographic variation in thc 'Hoy' call
of
the
Bobwhite.
Auk
95:
85-94.
honishi.
M.
19h
3.
'I'he
role
of
auditory feedback in the control
of
vocalisations in thc domestic
k~wl.
Z.
Ticrpsychol.
20:
349-367.
I\rondsma.
D.E.
1982.
Learning and the ontogeny
of
sound signals
in birds.
In
Kroodsma.
D.E.
(ed.) Accoustic Communication in
13irds.
Vol.
2:
1-23.
New York Academic
Press.
Lade,
H.I.
&
Thorpe,
W.H.
1964.
Dove songs
as
innately coded
patterns
ol'
specific behaviour.
Nature
202:
366-368.
hlarler.
1'.
1970.
A
comparative approach to vocal learning: Song
development in White Crowned Sparrows.
J.
Comp. Physiol.
71:
1-25,
McGregor.
l?K.
K.
Thompson. D.B.A.
1988.
Constancy and change
in local dialects
ol'
the Corn Bunting. Ornis Scand.
19:
15
3-1
59.
Bthol.
F:col.
Evol.
3:
113-126. Mundinger.
P.C.
1982.
Microgcographic and macriigeographic
variation in the aquired vocalisations
ol
birds.
fri
Kroodsma.
D.E.
(ed.) Acoustic Communication in Birds.
Vol.
2:
147-200.
New
York: Academic
Press.
Nottebohm.
F.
1975.
Continental
patterns
ol'
song variability in
Zoriotrichia
coprrisis:
Somc
possible ecological c'nrrclatcs.
Am.
Nat.
Nottebohm.
F.
c1.
Nottebohm.
M.
197
I. Vocalisations and brceding
hehaviour
of
surgically dcalened King
DOWS
(Strrp~opdin
rismicc).
Anim. Behav.
19:
31
3-327.
Payne.
R.H.
K.
l'ayne.
L.1,.
199
3.
Song copying
arid
cultural trans-
mission in Indigo Buntings. Anim. Behav. 46: 1045-1065.
Soltal.
R.
&
Kohlf',
F.1.
1981.
Biometry.
New
Yurk: W.H. Ikeman.
Thorpc. W.H.
19
58.
The learning
of
song
patterns by birds. with
special reference
to
the song
of
thc Chalfinch.
Friryilla
rocMs.
Ibis
100:
535-570.
Williams.
J.M.
&
Slater. P1.B.
199
3.
Does Chaffinch
15.iiigilln
rodrbs
song vary with the habitat in which it is sung? Ibis
115:
202-
208.
109:
606-624.
