Biological Journal of the Linnean Society, 2002, 75, 21–26. With 2 figures
The geographical distribution of populations of the
large-billed subspecies of reed bunting matches
that of its main winter food
GIULIANO MATESSI*, MATTEO GRIGGIO and ANDREA PILASTRO
Dipartimento di Biologia, Università di Padova, via U. Bassi 58/B, I-35131, Padova, Italy
Received 28 January 2001; accepted for publication 7 September 2001
We investigated the effects of resource distribution on the population structure and distribution of a polytypic bird
species. We compared the presence of insect remains (mainly dormant larvae) in the winter diet of two reed bunting
subspecies, small billed E. s. schoeniclus and large billed E. s. intermedia and studied the distribution of this resource
within reed (Phragmites sp.) stems in seven north Italian localities where the two subspecies breed (three schoeniclus and four intermedia populations). We also tested if the distributions of winter insect resources and breeding
populations of the large bill subspecies overlapped. The distribution of the larvae in reed stems matched closely
the distribution of large billed breeding populations. The winter diets of the two subspecies were significantly
different in terms of frequency of insect remains. These results, when compared to theoretical models of parapatric distribution, suggest that the two subspecies may be subject to ecological (vicariant) selection maintaining
their reproductive isolation. © 2002 The Linnean Society of London, Biological Journal of the Linnean Society,
2002, 75, 21–26.
ADDITIONAL KEYWORDS: bill size – diet – ecological selection – Emberiza schoeniclus – insect larvae –
parapatry.
INTRODUCTION
The divergence of vertebrate populations, and birds in
particular, down the path that leads to speciation
through reproductive isolation is thought to occur
most often in allopatry (Lynch, 1989; Futuyma, 1997;
Grant & Grant, 1997; Grant et al., 2000), even though
at least two cases of possible sympatric speciation in
birds have been identified (Grant & Grant, 1989, 1997;
Smith, 1993). Geographically separated populations
can become reproductively isolated with time due to
development of pre–zygotic barriers to hybridization
through sexual selection or through natural selection
(ecological speciation). Mating behaviour barriers in
birds, especially involving song, are well documented
(Catchpole & Slater, 1995), examples of ecological
speciation due to the prevailing influence of either
phenology or habitat use are rare. Yet ecological, or
*Corresponding author. Current address: Zoological Institute,
Copenhagen University, Tagensvej 16, DK-220 Copenhagen
N, Denmark. E-mail: [email protected]
vicariant, selection is thought to be the most common
path of speciation (Futuyma, 1997). Examples of specific studies are those of the Galapagos finches, where
the distribution of different resources (seeds) determines the relationships among different sister species
and morphological variation within single species
(reviewed in Grant & Grant, 1997). In a recent study,
ecological variation was the best explanation for speciation patterns in the sharp-beaked ground finch
(Geospiza difficilis), even though this occurred on different islands of the Galapagos archipelago (Grant
et al., 2000). The African finch Pyrenestes ostrinus
represents a special case of intraspecific competition
leading to morphological divergence and sympatric
ecological isolation (Smith, 1990, 1993).
According to some mathematical models of resource
competition, patterns of parapatric distribution with
little or no overlap and/or hybridization among two
(sub)species can be determined by a few simple
factors: patchy habitats and a resource gradient for
the dominant (sub)species (Bull & Possingham, 1995).
This means that such conditions may theoretically be
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26
21
22
G. MATESSI ET AL.
sufficient to maintain reproductive isolation upon secondary contact between diverging populations.
The reed bunting (Emberiza schoeniclus, Linnaeus
1758) subspecies complex represents a suitable case
for the study of ecological selection as a force in speciation. The species is highly polytypic, the subspecies
vary primarily in bill size (Snow, 1994; Byers et al.,
1995). During winter in the western Palearctic, the
southern, resident subspecies E. s. intermedia have
been reported to break reed (Phragmites sp.) stems
and extract dormant insect larvae, probably favoured
by their large and resistant bill (e.g. Stegmann, 1948;
Isenmann, 1990; C. Guzzon, pers. comm.). During
summer, both the large bill form and the northern,
migratory small bill subspecies E. s. schoeniclus are
insectivorous (Hillcoat, 1994a). The reed breaking
behaviour has only been observed in a few E. s. intermedia individuals and nothing is known about the
frequency of the behaviour in general. Furthermore,
there is no evidence of such resource use in the small
billed populations, which feed mainly on seeds in the
winter (Hillcoat, 1994a). Genetic differentiation
among populations is small, indicating their recent
divergence (Grapputo et al., 1998). The western
Palearctic populations are highly fragmented due to
the patchy distribution and often limited size of reed
beds (Brichetti & Cova, 1976; Finlayson et al., 1992),
the favourite habitat for reproduction and wintering
(Nicholson, 1994). In the southern part of the breeding range, large and small billed populations breed
in sites situated close to each other (in some cases
within 50 km, Brichetti & Cova, 1976; Mezzavilla,
1989; Grapputo et al., 1998), apparently with very
little interbreeding (Grapputo et al., 1998). During
winter, however, northern, small billed individuals
reach the breeding grounds of the large billed reed
buntings and the two forms winter in sympatry (Pr ŷsJones, 1984; Amato et al., 1994). Attempts to identify
song differences between populations have demonstrated a weak influence of song as isolating mechanism (Matessi, 1999; Matessi et al., 2000a,b). In
particular, there is essentially no difference in the
response of males of either form to playback of the two
subspecies’ song (Matessi et al., 2000a; Matessi et al.,
in press). Here we investigate the potential role of
feeding ecology in explaining the actual distribution of
the two bill forms and, possibly, of their speciation
process by: (i) testing whether there is a difference in
frequency of insect components in the winter diets of
small and large billed populations; (ii) describing the
distribution of insect prey items in reed beds across
the boundary between the breeding range of the smallbilled form (schoeniclus) and the large-billed form
(intermedia); (iii) testing whether the geographical
distribution of large billed breeding populations is
related to the distribution of these resources.
METHODS
We collected samples of faeces of reed buntings of the
two subspecies, large billed intermedia (bill height
mean ± SE: males 6.72 ± 0.064 mm, N = 24; females
6.47 ± 0.086 mm, N = 18) and small billed schoeniclus
(bill height mean ± SE: males 4.94 ± 0.062 mm, N = 32;
females 4.97 ± 0.059 mm, N = 31), from individuals
mistnetted while feeding in reed beds of three different Italian localities (Fig. 1; Marano Lagunare
45°46¢N 13°09¢E; Punta Alberete 44°30¢N 12°16¢E;
Macchia Grande 41°45¢N 12°14¢E). Faecal samples
were collected between October and February in 1996,
1997 and 1998. The samples were collected by keeping
the birds for about 1 h in large paper envelopes placed
inside cotton bags used in ringing stations. Sex,
weight and bill size of the birds were measured at the
ringing stations. Females were assigned to the intermedia group if bill height was >5.9 mm, males if bill
height was >6.0 mm (for classification of subspecies,
see Amato et al., 1994). The envelopes were brought to
the laboratory and the samples were transferred
to plastic test tubes and frozen. We also collected
samples of insect larvae found inside reed stems for
comparison with faeces content.
We transferred 96 faeces samples to plastic Petri
dishes, dried the samples in a stove at 50 °C and
5
6
7
1
3
4
2
A
B
C
Figure 1. Distribution of the seven localities of reed stem
collection. The line separates the breeding areas of the
two forms. Solid circles: intermedia localities; open circles: schoeniclus localities. Numbers refer to the names of
localities given in Table 1. Open squares are localities of
faeces sample collection: A = Marano Lagunare; B = Punta
Alberete; C = Macchia Grande.
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26
WINTER DIET AND DISTRIBUTION OF REED BUNTINGS
weighed them. We then added a few drops of ethanol
(96%) in the Petri dishes to disaggregate the larger
lumps and separate the particles. Finally, the samples
were inspected under a dissecting stereoscope (¥40
magnification) to search for insect remains. In a few
cases, the particles were also mounted on a slide and
observed at a higher magnification (¥100).
During March 1997 we collected reed stems from
seven localities in northern Italy where one of the two
subspecies breed: fiume Sile (TV), Marano Lagunare
(UD), Busatello (VR) and Campotto (FE) for intermedia; Lago S. Croce (BL), Pusiano (CO) and Brabbia
(VA) for schoeniclus (Fig. 1). Within each locality, we
collected ten random stems from ten randomly chosen
locations (in total 100 stems at each locality). We measured total length and diameter at the base for 30 reed
stems (three in each ten-stem group, Table 1). We
inspected all reed stems for the presence of insect prey
items. All the leaves were stripped off and the stem
was opened from the top with a progressive incision of
1–2 cm. We collected all the insect prey items found
within reed stems. Larvae and pupae were classified
into order, dried in an oven at 60 °C for 24 h and
weighed. We identified in total two types of Diptera
larvae (Item Biomass (I.B.): type 1 = 0.025 g; type
2 = 0.025 g), two types of Diptera pupae (I.B. type
1 = 0.005 g; type 2 = 0.0075 g), two types of Lepidoptera
larvae (I.B. type 1 = 0.005 g; type 2 = 0.15 g) and one
type of Hymenoptera larvae (I.B. = 0.005 g).
Faeces and reed stem insects were not always
collected in the same locality because reed buntings,
although wintering in all of them, are present in some
in too little numbers to allow us to gather a useful
sample of faeces. At the same time, collecting reed
stem insects in the other localities did not serve our
purpose, these being too far from the contact zone.
RESULTS
We found insect remains in only three of 60 faeces
samples from schoeniclus individuals and in 32 of 36
23
samples from intermedia individuals (Fisher’s exact
test, P < 0.0001), the latter belonging to at least two
of the types of larvae found in the stems. Some of the
intermedia samples also contained the remains of
unidentified adult arthropods. Both schoeniclus and
intermedia samples contained seed remains and other
vegetable material.
Reed stem size did not differ between the localities
where the two forms breed, but there was high variability in reed sizes among localities within the same
form groups (nested ANOVA, N = 210: height: among
groups F1,5 = 0.162, NS, among subgroups F5,203
= 15.351, P < 0.001; diameter: F1,5 = 0.131, NS, among
subgroups F5,203 = 22.507, P < 0.001 (Table 1).
Localities within bill form groups did not differ in
number or biomass of insect prey items (intermedia:
N = 40, d.f. = 3; number, c2 = 1.751, P = 0.63; biomass,
c2 = 4.26 P = 0.23; schoeniclus: N = 30, d.f. = 2; number,
c2 = 0.154, P = 0.99; biomass, c2 = 0.531, P = 0.75;
Kruskal–Wallis exact test). On the contrary, there was
a significant difference of the number and biomass of
insect prey items between the localities grouped
according to bill form (Fig. 2). Items were more abundant and had a larger biomass in the localities where
intermedia breeds (N = 70, d.f. = 1; number, c2 = 7.326,
P = 0.007; biomass, c2 = 28.356, P < 0.001,
Kruskal–Wallis exact test). The tests were significant
also after sequential Bonferroni correction for using
the same data in two consecutive tests (initial a = 0.05,
k = 2; Rice, 1989). The maximum biomass of insect
prey items found in a schoeniclus locality was less
than half of the minimum biomass found in an intermedia locality (Fig. 2).
DISCUSSION
We found that: (i) insect remains were present almost
only in the winter diet of large billed intermedia reed
bunting populations wintering in Italy; (ii) the distribution of this food resource at the end of the winter
had a sharp geographical gradient; (iii) the distribu-
Table 1. Comparison of reed stem size in samples collected in the seven localities in northern Italy where the two reed
bunting subspecies breed. Means and standard deviations are given for the 30 reed stems in each locality. Numbers in
parentheses refer to the map in Figure 1
Breeding form
Locality
Co-ordinates
Height (± SD cm)
Large bill
Sile (1)
Marano La. (2)
Busatello (3)
Campotto (4)
L.S.Croce (5)
Pusiano (6)
Brabbia (7)
45°33¢N
45°46¢N
45°07¢N
44°36¢N
46°08¢N
45°49¢N
45°48¢N
176.2
134.2
192.8
219.2
198.8
151.5
162.2
Small bill
12°10¢E
13°09¢E
11°04¢E
11°51¢E
12°20¢E
9°17¢E
8°42¢E
±
±
±
±
±
±
±
20.8
37.2
47.8
59.3
57.2
46.1
27.8
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26
Diameter (± SD mm)
5.87
4.57
6.14
7.59
5.70
6.51
5.04
±
±
±
±
±
±
±
0.61
1.18
1.13
1.74
1.11
1.59
0.90
24
G. MATESSI ET AL.
Number of insect items
2.5
(A)
2.0
1.5
1.0
0.5
0.0
0.25
(B)
Item biomass (g)
0.2
0.15
0.1
0.05
0.0
-0.05
Sile
Marano Busatello Campotto
L.S.Croce Pusiano Brabbia
Locality
Figure 2. Mean and standard error of number of insect
prey items (A) and of biomass of insect prey items (B) found
within reed stems for each breeding locality. Solid circles:
intermedia localities; open circles: schoeniclus localities.
tion of insect prey items closely matches the breeding
distributions of the two subspecies. In particular,
insects were abundant only in localities where the
large-billed subspecies breeds. In nearby localities,
occupied during the breeding season by the smallbilled subspecies, insects in the reed stems were found
much less frequently and their maximum biomass half
of that found in localities of the large-billed form. Altogether, our data suggest an association between insect
abundance and the geographical distribution of the
two reed bunting subspecies.
Our results seem to fulfil the conditions that are
considered sufficient for little or no mixing of the populations according to theoretical models of parapatric
distribution. Particularly, Bull & Possingham (1995)
presented a model that explored the conditions for the
development of ecological parapatry, where taxa have
a narrow overlap zone but there is no hybridization
(Key, 1982). The model simulates the population
dynamics of two taxa, one ecologically dominant over
the other, but limited by a resource gradient, which
live in a heterogeneous environment formed by high
quality ridges and low quality valleys. The authors
concluded that these conditions are sufficient for the
development and maintenance of ecological parapatry
(Bull & Possingham, 1995). The two subspecies of reed
bunting we investigated seem to fit well in the ecological assumptions of parapatry dynamics. As the model
requires, the two forms differ in ecology, the large bills
having access to an extra resource, which has a strong
gradient in distribution. Moreover, the habitat is not
uniform for either of the forms, since reed beds are
scattered and of varying size and quality. Therefore we
suggest that the different ecology of the two forms
could be sufficient to maintain their isolation, with
other pre-mating mechanisms playing a minor role, if
any. The subspecies could be at an early stage of
parapatric ecological speciation, after having been
separated in glacial refuges with different resource
composition (Hewitt, 1996). The small differences in
song could be of secondary origin and depend mainly
on genetic and ‘cultural’ drift (Matessi, 1999; Matessi
et al., 2000b, 2001).
However, some of the conditions required by the
model need further investigation. While we can expect
large billed reed buntings, which also have a larger
body (Hillcoat, 1994a; C. Guzzon, pers. comm.), and
possibly are in better general condition during winter
due to a diet richer in protein, to be dominant over
small billed individuals, direct competition and dominance relationships between the subspecies have not
yet been demonstrated. A possible indication of competition could come from the evidence on territorial
responses of large billed males to playback of small
billed song (Matessi, 1999). On the other hand, the
Bull and Possingham model showed that a heterogeneous environment required lower strength of ecological interaction between the taxa, compared to
a homogeneous environment (Bull & Possingham,
1995).
Data on the frequency of mixed breeding populations and hybridization between the two forms are also
scarce (Brichetti & Cova, 1976). We are only aware of
a few large billed individuals captured at Brabbia,
a locality inhabited by small-billed individuals (D.
Rubolini, pers. comm.), and one small billed female
with a brood patch captured in spring at Busatello, a
locality where the large billed subspecies is present
(M. Pesente, pers. comm.). The existence of hybrids
cannot therefore be excluded, in which case hybridization parapatry could be playing a role, as, e.g. in some
North American warblers or in European flycatchers
(Key, 1982; Sætre et al., 1999; Pearson, 2000).
However, data on bill size from the north Italian
populations suggest that the two subspecies are
clearly differentiated (Grapputo et al., 1998).
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26
WINTER DIET AND DISTRIBUTION OF REED BUNTINGS
Sampling of faeces and reed stem insects in different localities should not undermine the results,
given that the difference in diet between the two
forms, wherever acquired, could suffice to produce
differences in the timing of migration to the breeding
territories or in fertility patterns. Furthermore, we
collected reed stem insects at the end of winter, when
the populations begin to reach their breeding ranges
(Hillcoat, 1994b) and are therefore still available to
intermedia individuals. Nonetheless, more exact
data on diet differences among breeding localities
could add useful detail to the ecological dynamics
of these populations.
In conclusion, our data revealed a close association
between the geographical distribution of insects
within reed stems and that of the reed bunting subspecies. Whereas this co-occurrence may explain the
lack of interbreeding between the two parapatric reed
bunting populations, other possible mechanisms of
morphological divergence should not be overlooked.
For example, the genetic mechanism of bill size determination could depend on a single locus, as found in
Pyrenestes ostrinus (Smith, 1993), and be sufficient to
explain the low overlap of bill size frequency distributions (Amato et al., 1994). A better knowledge of the
genetic basis and heritability of bill morphology is
required for a more general understanding of population differentiation dynamics, speciation and evolution
in this species.
ACKNOWLEDGEMENTS
We are particularly indebted to Carlo Guzzon (Ris.
Nat. Reg. ‘Foci dello Stella’, Friuli Venezia-Giulia),
Enzo Savo and Stefano Volponi for collecting the
majority of the faecal samples and for kindly providing their ringing data. We also thank all those that
have helped at the reed collection sites. Thanks to
Diego Rubolini and Marco Pesente for their data on
breeding populations. Alessandro Minelli kindly
helped with insect order determination. We are grateful to Peter R. Grant, Åke Berg, Robert Prŷs-Jones and
two anonymous referees for helpful comments on
various versions of the manuscript. We warmly thank
Guglielmo Marin, without whose guidance, suggestions and comments this work would not have been
possible. This research was partially supported by
ex60% and COFIN99 MURST grants to AP.
REFERENCES
Amato S, Tiloca G, Marin G. 1994. Winter sympatry of
two Reed bunting (Emberiza schoeniclus) subspecies in the
Venetian lagoon. Avocetta 18: 115–118.
25
Brichetti P, Cova C. 1976. La situazione nidificatoria del
Migliarino di palude in Valpadana. Uccelli D’Italia 1: 28–31.
Bull MC, Possingham H. 1995. A model to explain ecological
parapatry. American Naturalist 145: 935–947.
Byers C, Olsson U, Curson J. 1995. Buntings and Sparrows.
A Guide to the Buntings and North American Sparrows.
Sussex: Pica Press.
Catchpole CK, Slater PJB. 1995. Bird Song: Biological
Themes and Variation. Cambridge: Cambridge University
Press.
Finlayson CM, Hollis GE, Davis TJ. 1992. Managing
Mediterranean wetlands and their birds. Proceedings of the
Grado Symposium. Slimbridge: IWRB Special Publication
no. 20.
Futuyma DJ. 1997. Evolutionary Biology, 3rd edn.
Sunderland, MA: Sinauer.
Grant BR, Grant PR. 1989. Evolutionary Dynamics of a
Natural Population: the Large Cactus Finch of the Galapagos. Chicago, IL: University of Chicago Press.
Grant PR, Grant BR. 1997. Genetics and the origin of bird
species. Proceedings of the National Academy of Sciences
USA 94: 7768–7775.
Grant PR, Grant BR, Petren K. 2000. The allopatric phase
of speciation: the sharp-beaked ground finch (Geospiza difficilis) on the Galapagos islands. Biological Journal of the
Linnean Society 69: 287–317.
Grapputo A, Pilastro A, Marin G. 1998. Genetic variation
and bill size dimorphism in a passerine bird, the reed
bunting Emberiza schoeniclus. Molecular Ecology 7:
1173–1182.
Hewitt GM. 1996. Some genetic consequences of ice ages, and
their role in divergence and speciation. Biological Journal of
the Linnean Society 58: 247–276.
Hillcoat B. 1994a. Reed bunting Emberiza schoeniclus: Food.
In: Cramp, S, Perrins, CM, eds. Birds of the Western Palearctic, Vol. IX. Oxford: Oxford University Press, 281–283.
Hillcoat B. 1994b. Reed bunting Emberiza schoeniclus:
Breeding. In: Cramp, S, Perrins, CM, eds. Birds of the
Western Palearctic, Vol. IX. Oxford: Oxford University Press,
287–289.
Isenmann P. 1990. Comportement alimentaire original chez
le bruant de roseaux, E. s. witherbyi, sur l’ile de Majorque.
Nos Oiseaux 419: 308.
Key KHL. 1982. Species, parapatry and the morabine
grasshoppers. Systematic Zoology 30: 425–458.
Lynch JD. 1989. The Gauge of Speciation. In: Otte, D, Endler,
JA, eds. Speciation and its Consequences. Sunderland, MA:
Sinauer, 527–553.
Matessi G. 1999. Evolutionary patterns in European
populations of the reed bunting (Emberiza schoeniclus ssp.).
Unpublished PhD Thesis, University of Ferrara.
Matessi G, Dabelsteen T, Pilastro A. 2000a. Responses to
playback of different subspecies songs in the Reed Bunting
Emberiza s. schoeniclus. Journal of Avian Biology 31:
96–101.
Matessi G, Dabelsteen T, Pilastro A. (In press). Subspecies
song recognition in a Mediterranean population of the reed
bunting Emberiza s. intermedia. Italian Journal of Zoology.
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26
26
G. MATESSI ET AL.
Matessi G, Pilastro A, Marin G. 2000b. Variation in quantitative properties of song among European populations of
the reed bunting (Emberiza schoeniclus) with respect to bill
morphology. Canadian Journal of Zoology 78: 428–437.
Matessi G, Pilastro A, Marin G. 2001. Population memetic
analysis of variation of song, geographical distribution and
bill morphology in the reed bunting. Selection, in press.
Mezzavilla F. 1989. Atlante Degli Uccelli Nidificanti Nelle
Province Di Treviso E Belluno (Veneto), 1983–88. Montebelluna: Museo Civico di Storia e Scienze Naturali.
Nicholson EM. 1994. Reed bunting Emberiza schoeniclus:
Habitat. In: Cramp, S, Perrins, CM, eds. Birds of the Western
Palearctic, Vol. IX. Oxford: Oxford University Press,
278–279.
Pearson SF. 2000. Behavioural asymmetries in a moving
hybrid zone. Behavioural Ecology 11: 84–92.
Pr ŷs-Jones RP. 1984. Migration patterns of the reed bunting,
Emberiza schoeniclus schoeniclus, and the dependence of
wintering distribution on environmental conditions. Gerfaut
74: 15–37.
Rice WR. 1989. Analyzing tables of statistical tests. Evolution
43: 223–225.
Sætre G-P, Král M, Bures S, Ims RA. 1999. Dynamics
of a clinal hybrid zone and a comparison with island hybrid
zones of flycatchers (Ficedula hypoleuca and F. albicollis).
Journal of the Zoological Society of London 247: 53–64.
Smith TB. 1990. Resource use by bill morphs of an African
finch: evidence for intraspecific competition. Evolution 71:
1246–1257.
Smith TB. 1993. Disruptive selection and the genetic basis
of bill size polymorphism in the African finch Pyrenestes.
Nature 363: 618–620.
Snow DW. 1994. Reed bunting Emberiza schoeniclus: Distribution and Population. In: Cramp, S, Perrins, CM, eds. Birds
of the Western Palearctic, Vol. IX. Oxford: Oxford University
Press, 279.
Stegmann B. 1948. Functional significance of subspecific
characters among Reed Buntings (Emberiza schoeniclus L.).
Zoologische Zhurnal 27: 241–244 [in Russian].
© 2002 The Linnean Society of London, Biological Journal of the Linnean Society, 2002, 75, 21–26