Content uploaded by Stefan Pentzold
Author content
All content in this area was uploaded by Stefan Pentzold on Dec 19, 2014
Content may be subject to copyright.
ORIGINAL ARTICLE
Breeding habitat preference and foraging of the Cyprus Wheatear
Oenanthe cypriaca and niche partitioning in comparison
with migrant Oenanthe species on Cyprus
Christoph Randler Æ Constanze Teichmann Æ
Stefan Pentzold
Received: 10 March 2009 / Revised: 23 April 2009 / Accepted: 8 June 2009 / Published online: 8 July 2009
Ó Dt. Ornithologen-Gesellschaft e.V. 2009
Abstract Niche partitioning has been examined in
breeding bird communities and in winter quarters, but has
received less attention when comparing a resident breeder
and migrants during spring. Here, such an assemblage of
species of the same genus (Oenanthe) and guild were
analysed on the Mediterranean island of Cyprus. Northern
Wheatear O. oenanthe and Eastern Black-Eared Wheatear
O. hispanica melanoleuca were migrants, while Cyprus
Wheatear O. cypriaca was resident. Migrant wheatears
were more common in open habitats without trees, with a
lower proportion of vegetation in the 10 cm layer, less tree
and bush cover, a higher proportion of herbaceous layer,
and a higher amount of bare areas. By using a discriminant
function, we found that O. oenanthe was least tolerant
towards a high proportion of bush/tree cover, and O. cyp-
riaca was most tolerant, with O. h. melanoleuca in
between. Also, O. oenanthe tolerated the least proportion
of vegetation in the lowest layer, and O. h. melanoleuca the
highest. O. oenanthe hunted more often by hop-and-peck
and O. cypriaca more often used sallying and perch-and-
pounce. O. oenanthe was the most ground-dwelling species
with low perch heights and highest number of hops per
minute and hops per movement, while O. cypriaca was the
most arboreal species with the highest perches. Mean for-
aging rate did not differ between the species. A principal
component analysis followed by a discriminant function
showed that O. cypriaca has a high amount of aerial sal-
lying and perch-pounce hunting behaviour with fewer
hops, while O. oenanthe represents the contrary with hop-
and-peck movements on the ground and with fewer flights.
The data further indicate a clearer separation between O.
oenanthe and O. cypriaca while O. h. melanoleuca lies in
between utilising both foraging modes.
Keywords Breeder–migrant assemblage Coexistence
Foraging mode Breeding habitat preference
Niche partitioning
Introduction
Niche partitioning has been most often examined in
breeding bird communities (Cody and Walter 1976; Cody
1985), and to a smaller extent in the Palearctic–Afrotropical
migrant system or in the Neotropics, where winter visitors
compete with residents during the northern winter (Green-
berg 1995; Sherry and Holmes 1996; Salewski et al. 2003).
However, niche partitioning between a resident breeder and
migrants during spring or autumn migration has received
less attention (see, e.g., Herrera 1978; Hutto 1985; Green-
berg 1986). Nevertheless, during migration in Europe,
many individuals of many species stopover on small Med-
iterranean islands like Cyprus and may present a consider-
able source of competitors to breeding bird species. Food
may be the most important ecological factor in such
migrant–resident assemblages (Sherry and Holmes 1996;
Salewski et al. 2003). Here, breeding versus migrating
species of the same genus (Oenanthe) were analysed to
Communicated by F. Bairlein.
C. Randler (&)
University of Education Heidelberg,
INF 561-2, 69120 Heidelberg, Germany
e-mail: randler@ph-heidelberg.de
C. Teichmann S. Pentzold
Faculty of Biology, Pharmacy and Psychology,
Institute of Biology II, University of Leipzig,
AG Tier- und Verhaltensphysiologie, Talstraße 33,
04103 Leipzig, Germany
123
J Ornithol (2010) 151:113–121
DOI 10.1007/s10336-009-0432-0
assess niche partitioning (Greenberg 1986) with a focus on
breeding habitat and foraging (see Baumann 2001; Salewski
et al. 2003). Previous work on migrants in comparison to
winter visitors found that migrants were more eurytopic and
exploited the more open parts of the habitats (Leisler 1992).
Also, foraging speed and rate was higher in migrants
(Leisler 1992), and Palearctic migrants were more flexible
in their foraging strategies and generally more opportunistic
(Leisler 1992). Migrants had a smaller ecological niche
compared to residents (Salewski et al. 2003). In detail,
shrike (Lanius) species in southern Africa differed in their
perch height and pouncing distances (Bruderer 1994),
migrant Golden Orioles Oriolus oriolus fed in higher strata
and more at the edges of trees in comparison to the resident
O. larvatus and O. auratus (Baumann 2001). During a year-
round observation, Herrera (1978) found that residents have
a narrower repertoire of foraging tactics than non-resident
species. In wheatears, Leisler et al. (1983) reported that the
migrant Oenanthe pleschanka foraged the quickest by using
flights of longer distances compared to resident species.
Here, we focus on species of the genus Oenanthe which
are more or less generalists, hunting arthropods opportu-
nistically, and have overlapping diets (Panov 2005). Recent
studies on some of the 22 species of this genus were carried
out in the breeding areas and focused on niche segregation
(Kaboli et al. 2006, 2007a) and on interspecific competition
and territoriality in the winter quarters in Kenya (Leisler
et al. 1983). Studies on Northern Wheatears O. oenanthe
during migration focused on fattening strategies and
optimal migration (Schmaljohann and Dierschke 2004;
Dierschke et al. 2005). No study has yet compared niche
partitioning of Oenanthe species during spring migration in
a breeder–migrant system. Wheatears seem favourable for
the study of competition since they are opportunistically
hunting songbirds that feed mainly on insects, form a true
guild because they generally exploit the same kind of
resources (overlapping diets) and they are easy to observe
(Leisler et al. 1983; Kaboli et al. 2006). Therefore, niche
partitioning should not be based on differences in food
itself because wheatears are generalist insectivorous spe-
cies (Potapova and Panov 1977; Panov 2005; Kaboli et al.
2006). However, there may be as yet undetected differ-
ences in the size of the prey and the percentage of distinct
taxa.
Migrant species in our study system were Northern
Wheatear O. oenanthe and Eastern Black-Eared Wheatear
O. hispanica melanoleuca and the resident species was the
Cyprus Wheatear O. cypriaca which has been treated as a
separate species since 1982 (Sluys and van den Berg 1982),
being formerly treated as a subspecies of the Pied Wheatear
O. pleschanka. Some of the previous studies mentioned
above also focused on Oenanthe species but were based on
breeding bird communities and did not cover O. cypriaca
(Kaboli et al. 2006, 2007a). As expected from morpho-
logical data (Kaboli et al. 2007b), O. cypriaca should be a
vegetation-tolerant species utilising flimsy perches and,
similar to O. pleschanka and O. h. melanoleuca, should use
aerial sallying and perch-pounce as foraging tactics.
Aims of this study were (1) to describe the breeding
habitat preference of O. cypriaca and how it differs from
migrating species of the genus Oenanthe, (2) analyse for-
aging behaviour in comparison to other Oenanthe species
to test the hypothesis that O. cypriaca differs from
migrating Oenanthe species, and (3) provide evidence for
niche partitioning by comparing a resident species with
migrants of the same genus and guild.
Methods
Species and study area
O. cypriaca has been treated as separate species since 1982
(Sluys and van den Berg 1982) because of morphometric
measurements, less pronounced sexual dichromatism and
significantly different song than O. pleschanka (Christen-
sen 1974; Sluys and van den Berg 1982; Bergmann 1983).
Behavioural observations and ecological studies are rare. A
few sentences were published about the habitat (Oliver
1990) without extensive quantitative research. In keeping
with Leisler et al. (1983) and Salewski et al. (2003), we
chose habitat and foraging as the relevant factors in our
analysis. O. cypriaca is the only Oenanthe species breeding
on Cyprus, and there are no records of O. pleschanka
(Flint and Stewart 1992), while O. h. melanoleuca, and
O. oenanthe are regular migrants.
Cyprus is situated in the eastern Mediterranean. The
island is approximately 225 km long and 100 km wide and
is the third largest island in the Mediterranean (9,250 km
2
)
after Sicily and Sardinia (Stagg and Hearl 1998). It lies less
than 100 km from Turkey, and less than 200 km from
Syria and Lebanon in the east Mediterranean basin at
34°33
0
–35°42
0
N, 32°16
0
–34°36
0
E (Jones 2006). The island
is dominated by two mountain ranges, the Troodos
mountains on the Greek side rising up to 1,961 m, and
further north, the Kyrenia range rising to 1,024 m on the
Turkish side of the island. Cyprus has an extreme Medi-
terranean climate with long, very hot, dry summers and
cool, wet, changeable winters (Flint and Stewart 1992).
The island has a variety of natural vegetation; 18% of the
island is woodland (Stagg and Hearl 1998; Jones 2006).
Fieldwork
Field work took place from 22 March to 21 April 2008 and
totalled[300 h. We assessed the breeding habitats within a
114 J Ornithol (2010) 151:113–121
123
radius of 100 m for O. cypriaca and control habitats with a
radius of 25 m in migrant Oenanthe species. This was
chosen because migrants often use only a very small area,
and measuring more than 25 m radius would have included
vegetation structure that was not utilised to a large extent.
Furthermore, preliminary observations in 2005 (by C.R.)
indicated that 100 m for breeding habitat and 25 m for
migrant Oenanthe seem a good approximation. Therefore,
some categorical variables could not be compared between
migrants and breeders (e.g. the number of trees or bushes
within a given category, e.g. between 1 and 3 m, see
below). Breeding habitats of O. cypriaca were defined as
places where an individual was present and was singing or
foraging. As the species is rather common on the island and
it does not migrate further north or breed anywhere outside
Cyprus, this definition is sufficient. Furthermore, O. cyp-
riaca males immediately occupy their breeding territories
after arrival from their winter quarters (personal observa-
tion). Control habitats (100 m radius) were assigned when
there was no wheatear present and, additionally, playback
experiments were used to ascertain the absence of
O. cypriaca because it responds strongly to conspecific
playbacks (Teichmann et al. 2008).
Foraging of the target species was observed whenever
possible (see Salewski et al. 2003).
Habitat variables
Inclination was assessed using four distinct categories: flat,
B10°, 11–20° and C21°. Exposition was assigned to one of
the following categories: N, NE, E, SE, S, SW, W and NW,
if possible. Relief dynamic was classed into ten groups:
B0.5 m, 0.51–1 m, 1.1–2 m, 2.1–3 m, 3.1–4 m, 4.1–5 m,
5.1–6 m, 6.1–7 m, 7.1–8 m and [8 m. We grouped the
breeding habitats roughly into nine groups by a first esti-
mation: bush land higher than 1 m, bush land lower than
1 m, pastures, set-aside fields, open plains, woodlands,
meadows, agricultural land and urbanised areas. To cate-
gorise the vegetation structure, we adapted ordinal catego-
ries from Dale and Manceau (2003): (1) open habitat
without trees, (2) open habitat with only a few widely
scattered trees/bushes, (3) tree/bush vegetation present but
less than 50% cover, (4) more than 50% covered by trees/
bushes, and (5) closed habitat with dense and continuous
cover of trees. Dominating tree height was classified
according to Dale and Manceau (2003), but two categories
(0 and 1) were added: (0) no trees/bushes, (1) most trees/
bushes B0.5 m, (2) most trees/bushes between 0.51 and
1.0 m, (3) most trees/bushes between 1.1 and 2 m, (4) most
trees between 2.1 and 4 m and (5) most trees[4 m. Further,
we assessed the number of trees in different height cate-
gories of B0.5 m, 0.51–1 m, 1.1–3 m and[3 m. According
to the approach of sigma-sociology (Schwabe and Mann
1990), abiotic or anthropogenic structures were estimated:
(1) the number of small stone piles (ordinal: 0, 1–5, 6–10,
11–15, 16–20, 21–25 and [25), and (2) anthropogenic
structures such as walls, post, stakes and fences (ordinal: 0,
1–5, 6–10, 11–15, 16–20, 21–25 and[25) in three different
heights: \1 m, 1–2 m and [2 m. Song post height was
measured as precisely as possible using a metre rule which
made it possible to assess heights up to 4 m relatively
precisely (2 m taller than a person), greater heights were
estimated using comparable objects and the method above.
Vegetation structure, i.e. density of layers, was assessed
in percentage in different heights: (1) tree cover, (2) bush
cover, (3) herbaceous layer (total), (4) herbaceous layer
above 4 cm, (5) herbaceous layer below 4 cm and (6)
general total vegetation cover at 10 cm height. Further, we
assessed the proportion of total bare area (Khlebosolov
et al.
2001), the proportion of bare ground (earth), and the
proportion of area covered with stones and rocks (Khleb-
osolov et al. 2001). Note that the total bare area might be
greater than the sum of bare earth and the area covered
with rocks and stones, because streets and other concrete
anthropogenic structures were also defined as bare areas.
Foraging variables
We observed foraging individuals that were not involved in
courtship display or in aggressive behaviour (Kaboli et al.
2007a). As O. cypriaca was sometimes singing on a song
post during the observations, we subtracted the time spent
singing from the feeding observation. Foraging techniques
were classified according to Dierschke (2003) and Kaboli
et al. (2006, 2007a): (1) ‘‘perch-and-pounce’’—the bird
scans the ground from a perch and attacks prey by sallying
out from the perch; (2) ‘‘hop-and-peck’’—the bird hops
forward, then pecks down at the prey or, if not pecking,
looks around and hops forward again; (3) ‘‘aerial sallying’’
or ‘‘flycatching’’—this technique is used to catch flying
insects in flight; and (4) ‘‘hovering’’—searching for prey
while in flight and capturing arthropods from a substrate
(twigs, leaves, ground). We further measured the distance
of a flight (e.g. performed during the aerial sallying or
during perch-and-pounce hunting), and counted the number
of hops and pecks (Dierschke 2003). Pecks indicate the
potential prey, as pecks are only performed when prey is
visible. Hops per peck were used as a measure of foraging
effort (Dierschke 2003). In most cases, it was impossible to
assess if a bird succeeded in its prey capture (see also
Moreno 1984). Also, mean hops per movement was
calculated as not all movements ended with pecking. We
measured the height of every perch, mostly directly by
using a ruler. Distances from perch base to landing point on
the ground were roughly estimated (sally distances;
Moreno 1984). Some variables were standardised on 60 s,
J Ornithol (2010) 151:113–121 115
123
i.e. number of aerial sallying, perch-and-pounce and hops
and pecks. Also, the number of hops per peck was used as
some kind of foraging efficiency. We excluded observa-
tions less than 30 s. Mean duration was 147.38 ± 11.87 s
(SE). Observations lasted from 30 to 920 s.
To get an overall variable of foraging rate, we stand-
ardised the data to foraging events per minute. Every
perch-pounce action, every aerial sallying event and every
peck in the hop-and-peck hunting was used as a hunting
event and standardised as foraging rate/minute.
Statistical analyses
We used chi-square statistics that were calculated either
from the raw data (e.g. when comparing different breeding
habitat types within a species) or from raw data organised
in a crosstab calculation. When the statistic indicated sig-
nificant differences, we checked the standardised residuals
to assess the differences. Standardised residuals of [2.0
were denoted P \ 0.05. For comparison of differences, we
used a t test or, if variables were ordinal scaled, a Mann–
Whitney U test. For the Principal Component Analysis
(varimax rotation), the variables were z-transformed. Data
were divided into subsets to avoid interference, e.g. when
comparing breeding and control habitats, and breeding and
migrant habitats, each subset was z-transformed separately.
Results
We sampled a total of 192 habitats, 129 O. cypriaca
breeding habitats, 17 control habitats (Oenanthe species
absent), and 46 habitats of migrants. Additionally, 13 song
posts were recorded without habitat analysis.
Comparison with control habitats
Inclination differed significantly between breeding and
control habitats. In detail, there were more control plots in
the category \10° and fewer above 20° (v
2
= 16.01,
df = 3, P = 0.001). Exposition did not differ between
breeding and control habitats (v
2
= 7.16, df = 7,
P = 0.412). Relief dynamic differed between breeding and
control habitats (v
2
= 21.49, df = 9, P = 0.01), with more
control habitats than expected in the category 0.51–1 m
and fewer than expected in the category [8 m. Concerning
the raw classification, control habitats were more often
found in agricultural land and less often in bush land
vegetation above 1 m (v
2
= 26.58, df = 8, P = 0.001).
The number of trees and bushes within categories differed
between breeding and control habitats only in the category
0.51–1 m. In this category, significantly more bushes were
present in O. cypriaca breeding habitats (Table 1).
Concerning anthropogenic structures and stone piles (see
Table 1), no differences between control habitats and
breeding habitats existed. Control habitats had more veg-
etation layer at 10 cm height, while breeding habitats were
characterised by a higher proportion of bush vegetation, a
higher percentage of herbaceous layer below 4 cm and a
higher proportion of rocks and stones (see Table 1).
Comparison with habitats of migrating Oenanthe
species
Exposition did not differ between breeding and migrant
habitats (v
2
= 2.39, df = 7, P = 0.935). Migrant wheate-
ars avoided bush land habitats [1 m, and preferred more
open habitats (v
2
= 79.09, df = 8, P \ 0.001). Migrant
wheatears were more common than O. cypriaca in open
habitats without trees, while O. cypriaca occurred more
often in breeding habitats where tree and bush vegetation
was present but with less than 50% cover (v
2
= 81.59,
df = 3, P \ 0.001; see Fig. 1). The dominating tree height
was lower in migrant habitats (Table 1, Fig. 2). The veg-
etation in different layers in migrant habitats (assessed as
percentage) is depicted in Table 1 and differed in almost
any variable measured from O. cypriaca breeding habitats.
Compared to breeding, migrant habitats are characterised
by a lower proportion of vegetation in the 10 cm layer, by a
lower tree and bush cover, a higher proportion of herba-
ceous layer in general and, especially in the layer below
4 cm, and a greater amount of bare earth (see analysis
below).
Comparison of breeding and control habitats
by principal component analysis
Four components were extracted with an Eigen-value [1.
PC1 explained 29.7% of the variance and represents the
openness of the vegetation in the low layer [herbaceous
[4 cm (0.96), total herbaceous layer (0.94), vegetation
cover at 10 cm (0.68), total bare area (-0.65) and rocks/
stone cover (-0.48)]. PC2 explained 21.2% of the vari-
ance and represents the bush and tree cover [bush/tree
cover (0.88), dominating tree height (0.68), proportion of
tree cover (0.80)]. PC3 explained 14.1% of the variance
and consisted of bare earth (0.80) and bush cover
(-0.73). PC 4 (10.3% of the variance explained)
containing a herbaceous layer below 4 cm (-0.83). The
results indicate that O. cypriaca avoids areas with too
much bush and tree cover, as well as areas with too little
cover (PC2) and areas with too much herbaceous layer
(Fig. 3). Control habitats are characterised by an extre-
mely high proportion of tree and bush cover or a rather
open habitat.
116 J Ornithol (2010) 151:113–121
123
Table 1 Comparison of different habitat variables between Oenanthe cypriaca (breeding habitat, 100 m radius), O. h. melanoleuca, O. oenanthe
(migrant habitat, 25 m radius) and control habitats (100 m radius) on Cyprus
Breeding
(n = 129)
a
Control
(n = 17)
a
Migrant
(n = 46)
a
Kruskal–Wallis
test (X
2
)
Breeding vs
control (Z)
Breeding vs
migrants (Z)
Trees/bushes 2.67 (0.05) 2.88 (0.30) 1.46 (0.08) 68.66*** -0.97 -8.39***
Dominating tree height 3.62 (0.09) 2.94 (0.36) 1.61 (0.25) 43.79*** -1.60 -6.55***
Tree category [3 m 3.82 (0.18) 3.47 (0.63) 0.45 (0.16) -0.50
Tree cat. 1.1–3 m 4.72 (0.15) 3.76 (0.57) 0.86 (0.23) -1.60
Tree cat. 0.51–1 m 4.15 (0.15) 2.17 (0.59) 1.06 (0.26) -3.29***
Tree cat. B0.5 m 3.79 (0.18) 3.00 (0.71) 1.69 (0.31) -1.03
Anthropogenic structures \1 m (cat.) 1.60 (0.21) 1.17 (0.48) 0.15 (0.13) -0.09
Anthropogenic structures 1–2 m (cat.) 1.75 (0.22) 2.17 (0.65) 0.13 (0.05) -0.99
Anthropogenic structures [2 m (cat.) 1.65 (0.21) 1.41 (0.60) 0.04 (0.03) -0.61
Stone piles (cat.) 3.55 (0.21) 2.94 (0.72) 2.45 (0.38) -1.31
Tree cover (%) 15.66 (1.06) 30.53 (8.57) 2.67 (0.84) 66.26*** -0.30 -8.25***
Bush/shrub cover (%) 24.21 (1.33) 16.00 (2.97) 8.22 (1.38) 45.60*** -2.21* -6.60***
Vegetation at 10 cm 42.60 (1.81) 53.88 (6.72) 26.96 (3.88) 22.07*** -1.88* -4.31***
Herbaceous layer (total) 29.98 (1.43) 38.35 (6.80) 44.07 (3.12) 16.38*** -1.01 -4.12***
Herbaceous [4 cm 26.15 (1.46) 36.59 (6.82) 30.43 (3.37) 1.72 1.20 -0.75
Herbaceous \4 cm 3.84 (0.46) 1.76 (0.86) 13.63 (2.34) 29.69*** -3.21*** -4.26***
Bare parts (total) 34.49 (1.31) 31.00 (4.82) 46.04 (3.24) 12.46** -
1.15 -3.26**
Bare earth 18.36 (1.02) 21.88 (4.93) 30.83 (3.17) 12.83** -0.01 -3.62***
Rocks/stones 12.92 (1.13) 7.47 (3.30) 13.80 (3.25) 9.94** 2.76** -2.14*
Kruskal–Wallis test (df = 2 for all comparisons), followed by pairwise Mann–Whitney U tests
*** P \ 0.001, ** P \ 0.01, * P \ 0.05
a
Means with standard errors in parentheses
Data from some categorical variables could not be compared because of different sample plot sizes (see ‘‘Methods’’ )
4321
Vegetation density
80
60
40
20
0
number of habitats
migrants
Cyprus Wheatear
Fig. 1 Comparison of habitats of O. cypriaca and of migrants. Four
main categories were assigned following Dale and Manceau (2003):
(1) open habitat without trees, (2) open habitat with only a few widely
scattered trees/bushes, (3) trees/bush vegetation present but less than
50% cover, (4) more than 50% covered by trees/bushes
543210
dominating bush/tree height
50
40
30
20
10
0
number of habitats
Migrants
Cyprus Wheater
Fig. 2 Dominating bush/tree height in O. cypriaca breeding and
migrant habitats according to Dale and Manceau (2003): (0) no trees/
bushes, (1) most trees/bushes B0.5 m, (2) most trees/bushes between
0.51 and 1.0 m, (3) most trees/bushes between 1.1 and 2 m, (4) most
trees between 2.1 and 4 m, (5) most trees [4m
J Ornithol (2010) 151:113–121 117
123
Niche partitioning between Oenanthe species
To assess niche partitioning, we carried out a discriminant
function based on a former principal component analysis.
The first factor PC1 explained 34.5% of the variance and
comprised total herbaceous layer (0.94), herbaceous layer
[4 cm (0.95), total bare area (-0.61), and vegetation
cover at 10 cm (0.61). The second factor PC2 explained an
additional 22.2% of the variance [proportion of tree cover
(0.78), bush/tree cover (0.83), dominating tree height
(0.69), stone/rocks (-0.60)] and represents cover and
height of trees. PC3 contains bush cover (0.68) and bare
earth (-0.85) and accounted for 13.5%. PC4 accounted for
9.3% of the variance [herbaceous layer \4cm(-0.91)].
The discriminant function based on the four factors of the
PCA revealed two functions, DF1 with an Eigen-value of
1.08 (90.8% of the variance explained) and DF2 with an
Eigen-value of 0.11 (9.2% explained variance). All four
factors contributed to the model in a stepwise procedure
(Wilks k = 0.432). PC2 (0.89) and PC3 (0.63) contributed
to DF1, and PC1 (0.43) and PC4 (0.80) to DF2, thus DF1
represents bush and tree cover and less open ground, while
DF2 represents the vegetation in the lower layer. In detail,
O. oenanthe was least tolerant towards a high proportion
of bush/tree cover and O. cypriaca was most tolerant,
with O. h. melanoleuca inbetween. With respect to DF2,
O. oenanthe tolerated the least proportion of vegetation in
the lowest layer and O. h. melanoleuca the highest (Fig. 4).
The cross-validation assigned 82.6% of all territories cor-
rectly. These analyses indicate that O. cypriaca is the most
arboreal species, while O. oenanthe prefers open habitats
with fewer trees and bushes and a vegetation layer between
0 and 10 cm. O. h. melanoleuca lies inbetween with a
higher tolerance for bushes and trees compared to
O. oenanthe, and a higher proportion of the vegetation
layer in the 0–10 cm strata compared to O. cypriaca.
However, the most striking fact in the explained variance
stems from the proportion of trees/bushes.
Foraging behaviour
The three species differed significantly with respect to their
main feeding techniques (Fig. 5). O. oenanthe hunted more
often by hop-and-peck (v
2
= 91.03, df = 2, P \ 0.001)
and O. cypriaca more often used sallying (v
2
= 17.08,
df = 2, P \0.001) and perch and pounce (v
2
= 73.49,
df = 2, P \0.001). O. h. melanoleuca did not differ from
the expected values (Fig. 5). O. oenanthe was the most
ground-dwelling species with low perch heights and
highest number of hops per minute and hops per movement
(Table 2), while O. cypriaca was the most arboreal species
with a greater amount of perch-and-pounce foraging and
aerial sallying and fewer hop-and-peck movements. Fur-
ther, O. cypriaca had the highest perches (Table 2). Mean
flight distance differed between O. oenanthe and both O. h.
melanoleuca and O. cypriaca, but not between O. h. mel-
anoleuca and O. cypriaca. Mean foraging rate did not
differ between the species.
To separate the species, we applied a principal compo-
nent analysis with varimax rotation on the z-transformed
scores of perch-pounce/min, aerial sallying/min, hops/min,
foraging rate/min, mean perch height and mean flight dis-
tances. Three factors were extracted with Eigen-values [1.
The first factor (33.9% of the variance explained) contained
aerial sallying (0.90) and mean perch height (0.85). The
3,02,01,00,0-1,0-2,0-3,0
principal component axis 1
4,0
3,0
2,0
1,0
0,0
-1,0
-2,0
-3,0
principal component axis 2
ControlBreeding
open ground
high cover in low layer
high bush/tree cover
Fig. 3 Principal component analysis of O. cypriaca breeding habitats
and control plots. Open circles Breeding habitats of O. cypriaca,
filled quadrates control area
20-2-4
Function 1
3
2
1
0
-1
-2
-3
-4
Function 2
3
2
1
group means
melanoleuca
oenanthe
cypriaca
Oenanthe species
high bush/tree cover
high herbaceous > 4 cm
open ground
Fig. 4 Discriminant function analysis of habitat preferences of
O. cypriaca, O. h. melanoleuca and O. oenanthe. Group centroids
are 0.545/-0.042 in O. cypriaca, -2.442/-0.646 in O. oenanthe and
-1.421/0.714 in O. h. melanoleuca
118 J Ornithol (2010) 151:113–121
123
second PC (23.0%) contained mean foraging rate (0.86)
and mean flight distance (-0.58). Finally, PC3 (17.1%)
was composed of perch-pounce (0.96) and hops/min
(-0.62). We subsequently used these components in a
stepwise discriminant function. All three components
contributed to the final model in the stepwise procedure
(Wilks k = 0.287). DF1 had an Eigen-value of 2.23, DF2
of 0.08. PC1 (standardised coefficient = 0.85) and PC3
(1.07) contributed to the DF1, composed of aerial sallying
and perch height (PC1) and of perch-and-pounce and hops
(PC3). DF2 was composed of PC2 (0.94) which is based on
mean foraging rate and mean flight distance. This shows
that O. cypriaca has a greater amount of aerial sallying
and perch-pounce hunting behaviour with fewer hops,
while O. oenanthe represents the contrary with hop-and-peck
movements on the ground and with fewer flights. The data
further indicate a clearer separation between O. oenanthe
and O. cypriaca while O. h. melanoleuca lies in between
utilising both foraging modes, from perch and from the
ground (Fig. 5). The cross-validation classified 74.6% of
the species correctly.
Discussion
This study provides data for the breeding habitat selection
and foraging of O. cypriaca, which is a poorly known
endemic species. The only comparable data from other
published work is the height of the song post (Oliver 1990).
This author noticed 37 song posts but with severe pseudo-
replication (Hurlbert 1984), at least one male contributed
11 times to his data. Half of the perches (59%) were between
5 and 10 m which is comparable to our study. Also, song
posts (5.88 m) were higher than hunting perches (2.8 m; see
also Oliver 1990). The breeding habitat of O. cypriaca as
found in this present study somewhat resembles that reported
in Kaboli et al. (2006) for O. pleschanka and O. h. mela-
noleuca with a high proportion of shrubs and trees. However,
in the direct comparison between O. h. melanoleuca and
O. cypriaca on Cyprus, O. cypriaca clearly is the most
arboreal species. Morphometric data (Kaboli et al. 2007b)
suggesting that O. cypriaca should use more arboreal
habitats than other species of the genus Oenanthe have
been confirmed in this study. One question may be that
O. cypriaca extends its habitat into the more open areas after
spring migration because competition is absent. However,
this is not the case, and O. cypriaca remain in their territories
and only after breeding may move to higher altitudes due to
high temperatures (Flint and Stewart 1992).
The discriminant analysis showed that O. h. melanoleuca
lay inbetween the other two species, because it forages
420-2-4
Discriminant Function 1
4
2
0
-2
-4
Discriminant Function 2
Group means
oenanthe
melanoleuca
cypriaca
Oenanthe species
high perch, sallyinghop/peck on ground
high foraging rate, low flight
distances
Fig. 5 Foraging differences between three Oenanthe species based
on a discriminant function analysis. Group centroids are 1.715/0.115
in O. cypriaca, -1.699/0.229 in O. oenanthe, -0.423/-0.453 in O. h.
melanoleuca
Table 2 Comparison of foraging behaviour of O. cypriaca, O. h. melanoleuca and O. oenanthe on Cyprus
O. cypriaca
a
O. h. melanoleuca
a
O. oenanthe
a
Kruskal–Wallis
test (X
2
)
cypriaca vs
melanoleuca (Z)
cypriaca vs
oenanthe (Z)
oenanthe vs
melanoleuca (Z)
Hops/peck – 7.53 (1.27) 11.38 (1.21) 3.29 ns – – -1.81 ns
Hops/movement – 4.45 (0.43) 7.07 (0.42) 15.05*** – – -3.88***
Hops/min 0.00 (0.00) 15.13 (3.52) 39.10 (4.09) 85.18*** -6.48*** -8.82*** -4.32***
Pecks/min – 3.47 (0.73) 4.26 (0.60) 0.67 ns – – -0.82 ns
Aerial sallying/min 0.91 (0.23) 0.18 (0.08) 0.04 (0.02) 19.10*** -2.84** -3.90*** -0.83 ns
Perch-pounce/min 2.38 (0.24) 0.59 (0.12) 0.00 (0.00) 75.41*** -5.23*** -7.96*** -4.91***
Perch height 2.85 (0.40) 0.57 (0.10) 0.06 (0.01) 87.82*** -5.70*** -8.19*** -5.88***
Foraging rate/min 3.30 (0.28) 3.00 (0.48) 4.20 (0.59) 3.61 ns -1.72 ns -0.46 ns -1.58 ns
Flight distance 4.36 (0.89) 7.09 (2.08) 1.16 (0.62) 52.53*** -0.76 ns -6.93*** -5.68 ***
Kruskal–Wallis test used for comparison, followed by Mann–Whitney U tests
a
Means with standard errors in parentheses
J Ornithol (2010) 151:113–121 119
123
sometimes similar to O. cypriaca by perch-pounce and
flights, but also in more open habitats by hop-and-peck
similar to O. oenanthe. O. cypriaca feeds mainly by perch-
and-pounce and by aerial sallying. However, it seems that at
least some of the foraging techniques are weather-depen-
dent in O. cypriaca (Teichmann et al. 2008), but further
work is needed to fully support this conclusion. Sex dif-
ferences have been reported in foraging O. h. hispanica, but
our data at present suggest no differences in O. cypriaca.
Sexual dimorphism is very low in O. cypriaca (Christensen
1974), which may be one of the reasons. However, a closer
look at sex differences should be done in further work.
As O. pleschanka has not been reported on Cyprus (Flint
and Stewart 1992), we compared foraging behaviour with
data from the literature: O. cypriaca had higher perches
(2.88 m) than O. pleschanka (which were about 1 m
reported in Snow and Perrins 1998; and 0.3–1.5 m in Glutz
von Blotzheim and Bauer 1988). Further, flight distances
were similar with 4–5 m in O. pleschanka (Glutz von
Blotzheim and Bauer 1988). Similar to O. cypriaca, O.
pleschanka spend less time on the ground and more time on
perches compared to other wheatears (Leisler et al. 1983).
O. pleschanka is defined as a perch-pounce species by
Kaboli et al. (2006), while O. h. melanoleuca showed more
aerial sallying (Kaboli et al. 2006).
Similar to the Oenanthe migrant–breeder community in
Africa observed by Leisler (1992), migrating Oenanthe
species preferred the more open areas (Leisler 1992).
However, foraging speed was similar which is in contra-
diction to Leisler et al. (1983). However, Salewski et al.
(2003) made some cautionary comments about the gener-
alisation of such studies, and it seems that examining
migrant–breeder communities might be an interesting field
for further research. Despite some overlap in foraging, both
O. h. melanoleuca and O. cypriaca are separated by dif-
ferent habitat requirements. Interspecific competitions (as
measured in terms of direct encounters) were rare, only
about five were observed during our field work. Direct
interspecific competition in Oenanthe seems more common
in the winter quarters (Leisler et al. 1983). A migrant–
breeder community may be less stable compared to a
breeder–winter quarters community. This may simply
result from the fact that winter visitors may stay longer
within the same area (Leisler et al. 1983), while migrants
may stay only for a short period of time (some hours up to a
few days), especially during spring migration (Dierschke
et al. 2005). These data further indicate that migrant–
breeder assemblages should receive further attention to
look at the coexistence of breeding species with competi-
tors during migration. As Salewski et al. (2003) empha-
sised, assessing coexistence in a migrant–breeder
community is far from simple but should be worth the
effort. This study adds to our knowledge of migrant–
breeder communities during spring migration and shows
how species occupy different niches and differ in habitat
selection and foraging.
Zusammenfassung
Bruthabitatpra
¨
ferenz und Nahrungssuche beim
Zypernsteinschma
¨
tzer Oenanthe cypriaca im Vergleich
zu durchziehenden Oenanthe-Arten auf Zypern
Die o
¨
kologische Einnischung und Koexistenz zwischen
nahe verwandten Vogelarten im Brutgebiet ist recht gut
untersucht, doch fehlen Untersuchungen besonders zur
Zugzeit. In dieser Arbeit untersuchen wir speziell eine
solche Vogelgemeinschaft wa
¨
hrend des Fru
¨
hjahrszuges.
Dabei wurden drei Arten von Steinschma
¨
tzern der Gattung
Oenanthe und aus derselben Gilde wa
¨
hrend des Aufent-
haltes auf der Mittelmeerinsel Zypern untersucht.
Wa
¨
hrend der Steinschma
¨
tzer O. oenanthe und der
Balkanschma
¨
tzer O. h. melanoleuca Zugvo
¨
gel sind, ist der
Zypernsteinschma
¨
tzer ein verbreiteter Brutvogel auf der
Insel. Im Fokus der Studie standen Habitat und Nah-
rungssuche. Die beiden Zugvogelarten waren ha
¨
ufiger in
offenen Habitaten ohne Ba
¨
ume anzutreffen, bei einer ge-
ringeren Vegetation in 10 cm Ho
¨
he und bei weniger
Deckung von Busch- und Baumschicht. Durchzu
¨
gler bevor-
zugten offene Fla
¨
chen (offene Erde/Sand), wa
¨
hrend in
den Habitaten des Zypernsteinschma
¨
tzers mehr Fels und
anstehendes Gestein zu finden waren. Eine Diskriminanz-
funktion zeigte, dass O. oenanthe die geringste Toleranz
gegenu
¨
ber einer hohen Deckung durch Bu
¨
sche und Ba
¨
ume
aufwies, wa
¨
hrend O. cypriaca die gro
¨
ßte Toleranz aufwies
und O. h. melanoleuca sich zwischen beiden Arten befand.
O. oenanthe jagte bevorzugt mit einer Hu
¨
pf-Pick-Technik,
O. cypriaca zeigte meistens Wartenjagd und Schna
¨
pper-
jagd. Von allen drei Arten war O. oenanthe sowohl vom
Habitat als auch der Nahrungssuche die typische Bodenart,
wa
¨
hrend O. cypriaca am ehesten eine arboreale Lebens-
weise zeigte.
Acknowledgments The study was supported by the Forschung-
skommission of the Deutsche Ornithologen-Gesellschaft (DO-G). Dr.
Susanne Rohrmann helped with the identification of the plant species.
Christian Vollmer and two reviewers commented on a previous draft
and gave helpful comments to improve the manuscript.
References
Baumann S (2001) Observations on the coexistence of Palearctic and
African Orioles Oriolus spec. in Zimbabwe. Vogelwelt 122:67–
79
Bergmann HH (1983) Some peculiarities of Cyprus bird voices.
Cyprus Ornithol Soc Rep 8:41–54
120 J Ornithol (2010) 151:113–121
123
Bruderer B (1994) Habitat and niche of migrant red-backed shrikes
in southern Africa. J Ornithol 135:474–475. doi:10.1007/
BF01639998
Christensen S (1974) Notes on the plumage of the female Cyprus Pied
Wheatear Oenanthe pleschanka cypriaca. Ornis Scand 5:47–52.
doi:10.2307/3675893
Cody ML (1985) Habitat selection in birds. Academic, London
Cody ML, Walter H (1976) Habitat selection and interspecific
interactions among Mediterranean sylviid warblers. Oikos
27:210–238. doi:10.2307/3543900
Dale S, Manceau N (2003) Habitat selection of two locally sympatric
species of Emberiza buntings (E. citrinella and E. hortulana).
J Ornithol 144:58–68. doi:10.1007/BF02465517
Dierschke V (2003) Stopover behaviour of Northern Wheatears
Oenanthe oenanthe on Helgoland: influence of foraging condi-
tions during autumn migration. Vogelwelt 124:165–176
Dierschke V, Mendel B, Schmaljohann H (2005) Differential timing
of migration in northern wheatears Oenanthe oenanthe: hurried
males or weak females? Behav Ecol Sociobiol 57:470–480.
doi:10.1007/s00265-004-0872-8
Flint P, Stewart P (1992) The birds of Cyprus. British Ornithologist’s
Union, Tring
Glutz von Blotzheim UN, Bauer KM (1988) Handbuch der Vo
¨
gel
Mitteleuropas. Band 11, Passeriformes 2. Teil. Turdidae. Aula,
Wiesbaden
Greenberg R (1986) Competition in migrant birds in the non-breeding
season. Curr Ornithol 3:281–307
Greenberg R (1995) Insectivorous migratory birds in tropical
ecosystem: the breeding currency hypothesis. J Avian Biol
26:260–264. doi:10.2307/3677328
Herrera CM (1978) Ecological correlates of residence and non-
residence in a Mediterranean passerine bird community. J Anim
Ecol 47:871–890. doi:10.2307/3676
Hurlbert SH (1984) Pseudo-replication and the design of ecological
field experiments. Ecol Monogr 54:187–211. doi:10.2307/
1942661
Hutto RL (1985) Habitat selection by nonbreeding, migratory
landbirds. In: Cody ML (ed) Habitat selection in birds.
Academic, London
Jones VR (2006) Comparative ecology of the endemic Cyprus
Warbler and the congeneric Sardinian Warbler: implications of
recent coexistence. PhD dissertation, University of Cambridge,
Cambridge
Kaboli M, Aliabadian M, Thevenot M, Guillaume CP, Prodon R
(2006) Ecological segregation between Iranian wheatears. Zool
Middle East 39:41–58
Kaboli M, Aliabadian M, Prodon R (2007a) Niche segregation,
behavioural differences, and relation to morphology in two
Iranian syntopic wheatears: the northern wheatear Oenanthe
oenanthe libanotica and mourning wheatear O. lugens persica.
Vie Milieu – Life Env 57:139–150
Kaboli M, Aliabadian M, Guillaumet A, Roselaar CS, Prodon R
(2007b) Ecomorphology of the wheatears (genus Oenanthe). Ibis
149:792–805. doi:10.1111/j.1474-919X.2007.00714.x
Khlebosolov EI, Lobov IV, Chabrova EA, Cheltsov NV (2001)
Structure of the ecological niche and geographical distribution of
northern wheatears (Oenanthe oenanthe, Turdina, Passerifor-
mes). Entomol Rev 81(Supp 2):226–231
Leisler B (1992) Habitat selection and coexistence of migrants and
Afrotropical residents. Ibis 134(Supp):77–82
Leisler B, Heine G, Siebenrock KH (1983) Einnischung und
interspezifische Territorialita
¨
tu
¨
berwinternder Steinschma
¨
tzer
(Oenanthe isabellinae, O oenanthe, O. pleschanka) in Kenia.
J Ornithol 124:393–413. doi:10.1007/BF01640360
Moreno J (1984) Search strategies of wheatears (Oenanthe oenanthe)
and stonechats (Saxicola torquata
): adaptive variation in perch
height, search time, sally distance and inter-perch move length.
J Anim Ecol 53:147–159. doi:10.2307/4348
Oliver PJ (1990) Observations on the Cyprus Pied Wheatear
Oenanthe pleschanka cypriaca. Sandgrouse 12:25–30
Panov EN (2005) Wheatears of Palearctic. Ecology, behaviour and
evolution of the genus Oenanthe. Pensoft, Sofia-Moskau
Potapova EG, Panov EN (1977) Structural patterns of jaw apparatus
in wheatears of the genus Oenanthe, with special reference to the
ways of decrease of food competition between closely related
species. Zool Zh 56:743–752
Salewski V, Bairlein F, Leisler B (2003) Niche partitioning of
two Palearctic passerine migrants with Afrotropical residents in
their West African winter quarters. Behav Ecol 14:493–502.
doi:10.1093/beheco/arg021
Schmaljohann H, Dierschke V (2004) Optimal bird migration and
predation risk: a field experiment with northern wheatears
Oenanthe oenanthe. J Anim Ecol 74:131–138. doi:10.1111/j.
1365-2656.2004.00905.x
Schwabe A, Mann P (1990) A method to describe an typify habitats of
birds, shown by example of the rock bunting (Emberiza cia).
Ecol Birds 12:127–157
Sherry TW, Holmes RT (1996) Winter habitat quality, population
limitation, and conservation of Neotropical-Nearctic migrant
birds. Ecology 77:36–48. doi:10.2307/2265652
Sluys R, van den Berg M (1982) On the specific status of the Cyprus
pied wheatear Oenanthe cypriaca. Ornis Scand 13:123–128.
doi:10.2307/3676198
Snow DW, Perrins CM (1998) The birds of the Western Palearctic,
concise edition. Oxford University Press, Oxford
Stagg A, Hearl G (1998) A birdwatching guide to Cyprus. Arlequin,
Essex
Teichmann C, Pentzold S, Randler C (2008) Erste Ergebnisse zum
Jagdverhalten, Habitat und Playbackexperimenten des
Zypernsteinschma
¨
tzers Oenanthe cypriaca. Vogelwarte
46:356–357
J Ornithol (2010) 151:113–121 121
123
