Professional Documents
Culture Documents
A SYNTHESIS OF FIELD
AND LABORATORY STUDIES
RICHARD A. ZANN
School of Zoology,
La Trobe University, Australia
OXFORD
VNIVERSITY PRESS
Reprinted 2002
A catalogue record for this book is available frolll the British Libraly
ISBN 0-19-854079-5
Preface
The Australian Zebra Finch is a popular cage bird in many countries, not
simply among hobbyists, but also among scientists where it is the pre-
ferred subject for laboratory research in a range of biological disciplines.
The primary aim of this book is to integrate these di verse laboratory
studies and place them in the context of the biology of the animals in the
wild so that a more complete picture of the adaptations and life history
of the species will emerge. From this 1 hope new understandings arise
that can act as catalysts for better research and lead, on the one hand, to
more biologically relevant questions by laboratory workers, and on the
other, to new insights into the range of adaptations field workers can
investigate. This synergistic interplay between field and laboratory stud-
ies is already producing results, for example, in the understanding of
paternity and mate choice (Chapters 9 and 11) and there is ample scope
for similar advances elsewhere.
Given the vast number of publications on domesticated laboratory
Zebra Finches, it is impossible to review all of them, or to keep up with
the continuous flow, so 1 have selected those that complement the field
studies and aid understanding of the species. 1 should say 'subspecies'
beca use almost all research has been conducted on the Australian sub-
species of Zebra Finch. Nothing is known about the field biology of the
Lesser Sundas subspecies, and only limited aspects of its vocal and repro-
ductive behaviour are known from captive birds. There is sorne fascinat-
ing work to be done on this subspecies; it will not only fill a void, but will
provide interesting reflections on the biology of the Australian subspecies.
For most of the last 30 years much of our knowledge of the behaviour
of the Zebra Finch in the wild was based on Klaus Immelmann's work.
Unfortunately, only a small proportion was published in English, in par-
ticular in his wonderful book, Australian Finches in Bush and Aviary
(1965). Readers of German had access to his massive papers from which
the English work was extracted, but if the citations in the literature are
any guide, these papers have been largely ignored. Therefore, another of
my'aims is to make accessible to readers of English additional details of
Immelmann's fine fieldwork. Of course, things have not stood still in
Australia since his visit and a number of ecological and behavioural stud-
ies have be en conducted on populations in different parts of the country
and 1 ha ve tried to set these in the context of Immelmann's contribution.
Jiro Kikkawa was responsible for my scientific introduction to Zebra
Finches when he brought sorne pairs into the teaching laboratory at the
University of New England where 1 was enrolled in Zoology in 1964.
Minuteness, neatness, wonderful colours and exquisite patterns were the
Vl Preface
impressions 1 remember. That same year 1 helped Jiro mistnet Zebra
Finches at his nearby field site, and two years later 1 began my Ph.D.
under his supervision at the University of Queensland on behaviour of
relatives of the Zebra Finch, namely, the Long-tailed, Black-throated and
Masked Finches. These are possibly more beautiful and charming than
'Zebbies' but ha ve less character. 1 gained insight into certain aspects of
the character of Zebra Finches through a relationship with a male Zeb-
bie 1 kept during the Ph.D. years. 'Fred' hatched in 1966 and was hand-
raised, consequendy he beca me sexually imprinted on me, mainly my
fingers, but also my face. In the aviary he would sometimes land on my
shoulders and back, but after a few drama tic episodes with the opposite
sex he was confined to a small cage on my filing cabinet for safety. With-
out fail he courted my finger when 1 placed it inside the cage. His song
and dance routine were species-typical, but punctuated with aggressive
pecks. He was renowned among my fellow postgraduates for his amus-
ing antics and his clockwork reliability: he always performed for visitors,
but only my fingers elicited it. Often on the first courtship of the day he
would mount my finger and copulate, occasionally leaving a droplet of
semen, but mostly he ended the routine with a vigorous peck ¡nstead.
Fred took a wife but they never reproduced despite building a nest.
They usually got on fine unless he happened to see me whereupon his
sexual preference would take over and he would court my face or fingers
then chase and attack his partner untíl 1 turned my back at which their
bond would be re-established and harmony restored.
Once 1 was unwittingly involved in his conditioning, or he in mine. He
had learned by trial and error to flick a droplet of water from his gravity
drinker onto the back of my neck when 1 was at my desk, a metre away.
When 1 turned around he would begin his courtship routine; presumably
my face or fingers or the courtship were reinforcing for him, and natu-
rally, his courtship was reinforcing for me. Poor Fred disappeared in
transit between Brisbane and Melbourne in 1972; he was in his sixth
year.
Zebra Finches are often taken for granted by Australians beca use they
are numerous, noisy and persistent, but many, including myself, admire
them beca use they typify the 'litde Aussie batder'-the small and
insignificant, that somehow succeeds by simply hanging on and enduring
the vicissitudes of the vast, harsh country of inland Australia.
Finally, for those readers interested in statistical evidence for state--
ments in the text 1 have only given details where the work has not been
published, or is not available in a thesis lodged in a university library.
Acknowledgments
Jiro Kikkawa, my teacher and supervisor, stimulated my love for birds
and íntroduced me to Zebra Finches and their scientific possibilities. The
Preface Vtl
Bundoora, Vic
March 1995 R.A.Z.
Contents
4 Feeding ecology 44
Diet 44
Foraging 50
Nesting activity and diet 55
Feeding flocks 56
Summary 59
8 Populations 139
Fluctuations in size of populations 139
Survivorship 142
Sex ratios 145
Dispersal and mobility 148
Summary 155
10 Vocalisations 196
Cal! repertoire 196
Distance Call 205
Structural variation 205
Development 207
Neural control 212
Individual recognition 212
Song 213
Structure 214
Song variation 221
Dirccted and Undirccted Songs 222
Functions of song 224
Contents Xl
Appendix 1 288
Appendix 2 289
Appendix 3 291
Appendix 4 293
References 301
'After travelling all day and singing about many adventures of the
honeyeaters and other bird ancestors, the Zebra Finch women called softly
to their children to rest; the children however kept their little song about the
Mal!ee-fowl.
"My little children with the red ochred noses,
My little children with the red ochred noses,
Come and sleep."
The children sang softly until they fel! asleep.'
The Zebra Finch first became known to science at the start of the 19th
century when it was collected on one of the earliest and most lavishly
funded voyages of scientific exploration yet to leave Europe. Nicolas
Baudin was commissioned by Napoleon in 1800 to carry out scientific
and geographical surveys of the coasts of Australia in the ships, 'Géo-
graphe' and 'Naturaliste'. After mapping the west coast of Australia the
ships visited Kupang, then a village in West Timor, from 18 August to
13 November 1801 and again from 30 April to 3 June 1803 during
which the Zebra Finch was collected by the expedition's naturalists, the
most well known of whom was the headstrong Francois Péron. Zebra
Finches formed part of an enormous collection of 100,000 zoological
specimens that made the hazardous voyage back to France and which
eventually formed the basis of Péron's vast work 'Voyage de découvertes
aux Terres Australes' published in 1816. Possibly, sorne live Zebra
Pinches survived the journey back to Paris because Louis Jean Pierre
Vieillot, the great taxonomist, made obscure references to behaviour in
his famous book Les Oiseaux Chanteurs (Volume 2: 1805 and 1809).
However, Vieillot did not publish the taxonomic na me (Fringilla guttata)
and the scientific description of the new species until 1817 (Nouveau
Dictionnaire d'Histoire Naturelle vol xxi, p. 223). Alfred Russell Wallace
(1864) also collected the Zebra Finch from Timor during his famous
zoological investigations of the region; he named it 'Amadina insularis'.
The Australian Zebra Finch was not described until 1837, when the
celebrated ornithologist, John Gould, received the first specimens
collected from the plains of central New South Wales. He named it
XlV Historical note
'Amadina castanotis', which means Chestnut-eared Finch (Synopsis.
Birds of Australia pt. 1 pi. 10, 1 Jan. 1837). In his Handbook to the
Birds of Australia (1865) Gould wrote, 'The Chestnut-eared Finch is one
of the smallest of the genus (family) yet discovered in Australia; it is also
one of the most beautiful, and in the chasteness of its colouring can
scarcely be excelled.'
References to Zebra Finches later appear in the accounts of the early
explorers and the pioneering settlers of the vast inland of Australia. The
strong attachment between Zebra Finches and surface water in the
deserts frequently indicated obscure sources of life-saving drinking water
to desperate explorers (Davidson 1905) and bushmen (Carter 1903;
Lindsay 1963). The pioneering ornith ologists, almost aH of them ama-
teurs with amazing qualities of endurance and persistence, found Zebra
Finches a conspicuous component of the avifauna of the arid and semi-
arid regions of the continent, not least, for the vast numbers that concen-
trated around the more persistent waterholes during the endemic
droughts that are so much a feature of the Australian inland. Early
names for the Zebra Finch, besides the Chestnut-eared Finch, indude
'Waxbills' and 'Diamond Sparrows'.
Zebra Finches, of course, were significant to the indigenous inhabi-
tants of Australia, the aborigines; in particular, those tri bes that inhab--
ited the more arid areas of central and western Australia. Occasionally,
Zebra Finches were a source of food, usually being hunted by children,
and naturally, during drought they were an important indicator of the
presence of surface water hidden away in small rock holes for any wan-
dering bands of nomads in strange country. Zebra Finches were among
the few species that did not foul the minute water supplies upon which
aborigines often depended for their survival, and were purposely allowed
access via small gaps in the rocks that covered sorne holes. Among sorne
tri bes of central Australia, droppings of Zebra Finches, when mixed with
herbs, were used medicinally for headaches and other pains (Winfield
1982).
Given its prominence in the landscape and its usefulness, many tri bes
coined unique names for the species, often based on onomatopoeic ren-
derings of the nasal distance caUs. Serventy and Whittell (1976) list seven
aboriginal names for Zebra Finches from among the various language
groups in Western Australia: 'chiaga', 'newmerri' or 'nee-murri' (= 'red--
nosed fellow'), 'yim-eye', 'neamoora', 'nyinnyinka', 'nyi-nyi' and 'nye-
nye'. In central Australia the Pitjanjatjara people called Zebra Finches
the 'nyi-nyi' (Isaacs 1980; Winfield 1982) or 'njinji' (Cleland 1931) after
the grass whose seed they often eat; the Aranda people call them 'nyinka'
or 'nienji' (Mountford 1976) and the Walpiri people, 'ithi' (Condon
1955) or 'jindjinmari' (Meggitt 1971).
The Zebra Finch also figured in the mythical traditions and everyday
expressions of sorne groups of aboriginal people. Thus, expressions for
Historical note xv
abundance would draw similes with a flock of Zebra Finches (Finlayson
1935). In the mountain devil ceremonies of the Aranda tribe, body deco-
rations mimicking the throat stripes of the Zebra Finch male are made
on the backs of men of the Zebra Finch totem (Mountford 1976).
Zebra Finches figured prominently in the ceremoniallife of the Pitjan-
jatjara people of the northwest corner of South Australia. In one of their
'dreamtime' stories, which describes the creation of the world, there is a
children's song which tells of a great journey undertaken by the numer-
ous Zebra Finch people (the 'Nyi-Nyi') who encounter the human ances-
tors of other species of birds as they traverse vast parts of the southern
deserts visiting important sacred sites until finally returning to their
home country bringing a valuable discovery, namely fire for cooking
(Isaacs 1980).
Most ornithologists who specialise on a particular species do not
rema in emotionally detached from their subjects, and this is often the
impetus for their research. For the 12-year-old Klaus Immelmann the gift
of a pair of Zebra Finches sealed his career for life. Henceforth, the
Zebra Finch was his 'bird of fate' ('Schicksalsvogel') (Immelmann 1970)
and his love for the species made it the focus of his research for more
than 30 years until his premature death at 52. His first major publication
on Zebra Finches was his doctoral thesis on the function of species-spe-
cific markings which he undertook at the University of Zurich (Immel-
mann 1959). This massive, and often ignored work, demonstrated a first
rate experimental approach, especially when one takes into account the
scientific genre of the time. Soon afterwards he spent almost a year in
Australia, all of it in the field, studying the behaviour and ecology of
grassfinches, the Zebra Finch in particular. He was a pioneer in many
respects and was one of the first ornithologists to spend a wet season in
northwest Australia. Somehow he seemed to find time to make observa-
tions on many other species of animals as weH, including the human
inhabitants, and published extensively on aH these subjects in the years
foHowing his return to Germany. His observations on the breeding
response of Zebra Finches to drought-breaking rains attracted world-
wide attention from ornithologists, although the opportunistic nature of
breeding in Zebra Finches had long been remarked upon by Australian
observers. Back in Germany he resumed his experimental investigations
into domesticated Zebra Finches focusing on the development of
behaviour. At Braunschweig and Bielefeld he made pioneering and semi-
nal contributions to two fields of research, sexual imprinting and song
learning, both of which are still being vigorously pursued at a number of
institutions around the world.
Although Australian Zebra Finches had been held as cage birds
in Europe before the 1850s (Neunzig 1965), the first scientific study of
the Zebra Finch was not until a century later when Desmond Morris
began his behaviour studies of aviary birds under the guidance of Niko
XVl Historical note
Tinbergen at Oxford University. Morris's much cited paper on reproduc-
ti ve behaviour of captive Zebra Finches, published in Behaviour in 1954,
is also memorable for its artistic illustrations, entertaining text, and
orthodox ethological interpretations. Klaus Immelmann and Desmond
Morris together were instrumental in making the domesticated Zebra
Finch the avian model of choice for many laboratory investigators in an
array of scientific disciplines around the world.
1 Systematics and phylogeny
Zebra Finches belong to the estrildine group of finches of which there are
approximately 124 species found mainly in the tropical and subtropical
parts of Africa, India, South-East Asia, Australia and Melanesia.
Estrildines are the only granivorous group of passerines endemic to Aus-
tralia and authorities recognise 18 species, of which the Zebra Finch is
the most widespread and numerous.
Taeniopygia ]
Stagonopleura
Poephila Poephilini
Neochmia (grassfinches)
Emblema
Erythrura
Lonchura
Heteromunia Lonchurini
--< Spermestes
Amadina J (mannikins)
Estrilda
Amandava Estrildini
Lagonosticta (waxbills)
Uraeginthus ]
Pytilia
Fig. 1.1 Systematic relationship of estrildine finches to genera and tribe. (After
Baverstock et al. 1991.)
Emblema pictum is not closely related to the other genera and is placed
in its own monotypic genus.
The Lonchurini tribe
Three of the five genera in the Lonchurini have indigenous species in
Australia. Erythrura, a supergenus, contains the Gouldian Finch Chloe ..
bia gouldiae and the Blue-faced Parrot-Finch Erythrura trichroa; a fur---
ther 10 species of Erythrura live in Malesia and Melanesia. The genus
Heteromunia has only one species, the Pictorella Mannikin Heteromunia
pectoralis, and the Lonchura genus has two species, the Chestnut·-
breasted Mannikin Lonchura castaneothorax and the Yellow-rumped
Mannikin Lonchura flaviprymna.
()
~
... .
.... .........
(\
',.. \J ."
... '.'
form to the present day. In the arid interior, dominated by desert and
grassland steppe, no fragmentatíon and subsequent speciation has
occurred to any extent due to the absence of geographic barriers (Keast
1961; Schodde 1982) although Cracraft (1986) postulates the existence
of inland barriers in pre-Pliocene epochs. High levels of nomadism in
arid species ensures a continuous distribution and thus militates against
isolation and speciation (Keast 1961).
Today most species of grassfinches live in savanna woodlands and
these have distinct northern ('Torresian' biogeographic region, Schodde
1982) and southeast ('Bassian') components that roughly correspond to
1 Systematics and phylogeny 9
tropical and temperate woodlands. The tropical northern eucalypt wood-
lands with their understorey of tall grasses form a broad belt across the
north of the continent. They are subdivided into northwestern (Kimber-
ley and Arnhemland) and northeastern (Cape York and northeast Aus-
tralia) components by a tongue of arid country at the head of the Gulf of
Carpentaria, which acted as a barrier to movement. According to Keast
(1958), subsequent geographic isolation has led to speciation in the case
of Poephila acuticauda and P. cincta and subspeciation in Poephila per-
sonata (personata and leucotis), Taeniopygia bichenovii (bichenovii and
annulosa) and Neochmia phaeton (phaeton and iredalei), but Boles
(1988) has recently merged iredalei into phaeton.
Five grassfinch species live in the eucalypt and acacia woodlands of the
eastern periphery. Poephila cincta and N. ruficauda originated most
likely from northern forms, whereas N. modesta and N. temporalis
probably have southern or eastern origins. Neochmia modesta prefers
drier less dense woodlands, whereas N. temporalis tolera tes a range of
habitats from sparsely vegetated woodland to moist, closed forests and
thickets (Immelmann 1965a). Keast (1958) concluded that an arid bar-
rier at the base of Cape York isolated northern and southern popula-
tions, which have differentíated to the level of subspecies in P. cincta
(atropygialis and cincta), N. ruficauda (clarescens and ruficauda) and N.
temporalis (minor and temporalis). There is a third subspecies of N. tem-
poralis in South Australia where the Coorong arid tract isolated loftyi
from temporalis (Keast 1958).
The Diamond Firetail Stagonopleura guttata is the fifth species of
grassfinch of the southeastern woodlands. It belongs to a genus with
southern origins adapted to moíst closed forests, but has made an adap-
ti ve shift to drier but still fairly densely wooded habitats often bordering
watercourses and has an extended distribution over much of the wood-
lands of the south and east. However, no subspecies have differentiated
due to the absence of arid barriers across its range. The Beautiful Firetail
S. bella from the southeast coast of the maínland and Tasmania and its
counterpart the Red-eared Firetail S. oculata from southwest corner of
Western Australia are the only species of grassfinches to specialise in
damp, thíck vegetation, in and adjacent to eucalypt foresto Keast (1961)
believes that they differentiated into eastern and western species when
isolated by the vast arid Nullabor zone during post-Pleistocene clímatic
changes.
Origin of the Lesser Sundas Zebra Finch
The Australian Zebra Finch is thought to have crossed to Timor from
the Kímberley regíon sometime during a Pleistocene glaciation when the
sea level dropped sorne 100-150 m thus extendíng the coastline to
within sorne 72 km of Timor (Mayr 1944 a, b). The high mountains of
west Timor may have been visible to birds blown out to sea by cyclones
10 THE ZEBRA FINCH
from the northern coast of Australia. This might explain why Zebra
Finches dispersed to Timor and why Timor estrildines (Amandava aman-
dava, Lonchura quinticolor, L. punctulata, L. molucca and L. fuscata)
that are well adapted to grassland and open country, failed to colonise
Australia: they could not see landfalls from coastal Timor.
The Zebra Finch is clearly the most superior disperser among the Aus-
tralian estrildines and this probably holds for most of the Indonesian
estrildines as well. Mayr (1944a,b) made the definitive zoogeographical
study of the transitional avifauna of Wallacea, the biogeographical zone
that separates the Oriental and Australian avifaunas and new data ha ve
been reviewed by White and Bruce (1986). On the one hand, Australia
received 20-22 species of birds of oriental origins via Timor, yet no
estrildines presently found on Timor are among the 10 species of passer-
ines that colonised Australia. On the other hand, Timor received 17 Aus-
tralian species of which seven are passerines including the Zebra Finch.
Of the 12 species of estrildines listed by White and Bruce (1986) for the
Lesser Sundas archipelago, the Zebra Finch has the most extensive distri-
bution by far (Chapter 2). Nonetheless, it has not crossed Wallace's line
and made it to Bali, with only three species of passerines of Australian
origin having done so.
petitors for the ancestral estrildines would have been ants, with parrats
and pigeons secondary competitors.
Evolution of the Australian climate
To understand the adaptations of the ancestral grassfinches and the pro-
cesses that led to their differentiation, it is necessary to appreciate the
changes to the Australian climate and enviranment in the late Tertiary
and early Quarternary periods (see Kemp 1981; Bowler 1982; Nix 1982;
Frakes et al. 1987; among others).
At the end of the Miocene epoch (10-5 million years ago) global cool-
ing formed high pressure mid-Iatitude systems which blocked moist mar-
itime air masses from entering the interior of Australia thereby causing
extensive desiccation (Bowler 1982; Frakes et al. 1987). There was a
pronounced, but gradual transition in vegetation fram widespread rain-
forest cover, to open forests and woodlands, to dry scrublands, and,
finally, to grasslands. Aridification continued into the mid- to late
Pliocene (Galloway and Kemp 1981) and reached its present state in the
Late Pliocene to Early Pleistocene, about 2 million years ago (White
1986).
Increased seasonality in moisture and temperature arose due to the
formation of a monsoonal climatic regime of summer rainfall and dry
winters. In southern Australia this later shifted to the opposite regime of
wet winters and dry summers. Nix (1982) believes that this seasonal defi-
ciency in water is the primary factor in the evolution of the Australian
biota found today and that these forces were greatest in the Tertiary and
relatively minor in the Quarternary. In contrast, Frakes et al. (1987),
postulate that the dramatic oscillations in humidity that began in the
Pliocene and intensified in the Pleistocene glaciations, especially in the
last 400,000 years when the frequency and amplitudes of the oscillations
were at their highest, have had the greatest ecological impact on the flora
and fauna of Australia. During the Quarternary, the Australian deserts
expanded on the northern and southern margins during glacial periods
but became so humid during the interglacial intervals that the arid belt
may have disappeared. At present the climate is about mid-cyde heading
fram a cool dry period for a more humid phase (Bowler 1982). It is rea-
sonable to condude that the dimatic extremes, especially that of aridity,
either of a seasonal or more permanent nature, caused severe ecological
stresses which have sifted the flora and fauna by means of differential
adaptation and extinction of species, thereby leaving but a fraction of the
Miocene biota to be encountered by humans on their arrival on the con-
tinent.
With this scenario in mind one can speculate on the type of di mate a
proto-grassfinch arriving' in northern Australia during the Plio-Pleis-
tocen e epochs might have experienced. While the dimatic data are more
sketchy than those for southern Australia, it is plausible to assume that
1 Systematics and phylogeny 13
grassland and shrubland habitats were expanding across northern Aus-
tralia as the continent became increasingly arid and seasonal. Appar-
ently, the present climate and vegetation resemble those of the late
Tertiary (Nix 1982). Thus, granivorous niches may have been largely
unoccupied and increasing in availability as new ones were created
around the time of arrival of the proto-grassfinch. The colonisers may
have even been pre-adapted to savanna conditions as Morley and Flen- .
ley's (1987) reconstructions indica te that Late Tertiary paleoclimates of
the Malay Peninsula had savanna corridors along which savanna species
could ha ve travelled to reach the stepping-stones of the Lesser Sundas
and the departure points for north-west Australia.
The first grassfinches to arrive in north-west Australia may not have
needed to make dramatic adaptive changes in order to survive and
disperse. From there they would have spread across the savanna wood-
lands of northern Australia and down the east coast. This widespread
ancestral species was subsequently subdivided by geomorphological and
ecological-climatic barriers into isolated areas where species differences
arose as forms adapted to more mesic conditions in sorne regions.
Speciation could have occurred repeatedly during the early glacial
periods of the Pleistocene when extreme aridification isolated patches of
humid forest in the south and east on the one hand, and savanna wood-
land in the north on the other. Cracraft (1986) has demonstrated a coin-
cidence of patterns of phylogenetic differentiation among the Poephila
grassfinches and the isolation of areas of endemism by barriers that
formed in the late Pliocene. During the Pleistocene many species must
have been driven to extinction. Species, such as the Zebra Finch and the
Painted Finch, which developed adaptations to the arid environments,
may have moved to the interior directly from the savanna woodlands of
the northwest. During periods of extreme aridity, Schodde (1982)
believes these two species took refuge in the mountain ranges of Central
Australia (MacDonnell, Tomkinson, Petermann, Musgrave and Everard
Ranges), Western Australia (Hamersley and Pilbara), South Australia
(Flinders Ranges) and Queensland (Selwyn Ranges) although the evi-
dence for this is unstated.
Summary
Two subspecies of Zebra Finches exist, the Lesser Sundas Zebra Finch
Taeniopygia guttata guttata (Vieillot, 1817) from eastern Indonesia, and
the Australian Zebra Finch Taeniopygia guttata castanotis (Gould, 1837)
from continental Australia. They belong to the Poephilini tribe of the
subfamily Estrildinae, family Passeridae. There are thirteen additional
species of Australian grassfinches in the Poephilini, of which the Double-
barred Finch Taeniopygia bichenovii is the only congener. The Zebra
Finch has a number of primitive phylogenetic characters, and may
14 THE ZEBRA FINCH
resemble the ancestral estrildine that colonised Australia from the Lesser
Sundas archipelago sometime during an earIy Pleistocene glaciation. The
ancient estrildines were in the vanguard of the 'northern elements',
mostly of African origin, that invaded Australia when it drifted north-
ward and contacted the Asian plateo The ancient estrildines radiated in
Australia and ha ve occupied most habitats and climates. This was aided
by climatic changes in Australia where increasing aridity extended grass-
lands and shrublands, the presumed habitat of the proto-grassfinches.
Among the old Australian endemics there are no passerine granivores
that might compete with proto-grassfinches, although pigeons, parrots
and ants would have provided sorne competition for seed. The Lesser
Sundas Zebra Finch is believed to ha ve differentiated when Timor was
colonised by Zebra Finches from northwest Australia during a late Pleis-
tocene glaciation when the water gap narrowed. When sea levels rose the
population was isolated from the mainland but was able to disperse
westward from Timor, reaching as far as Lombok. A detailed biochemi-
cal analysis of geographic variation in both subspecies of Zebra Finches
is needed.
2 Distribution and habitat
Fig. 2.1 Distribution (stippling) of the Zebra Finch in Australia and in the Lesser
Sundas archipelago of eastern Indonesia.
species of animals and it forms the southern limit of Wallace's line, the
boundary of major discontinuity in the distribution of Australasian and
Oriental avifaunas.
Climate and habitats of the Lesser Sundas
The Lesser Sundas lie in the rain shadow of northern Australia and have
the driest, most seasonal climate in Indonesia. They resemble the
wet-dry tropics of northern Australia in that they receive heavy mon-
soon rains from December to March but are completely dry for the rest
of the year (Nix 1976; White and Bruce 1986). Rainfall throughout the
archipelago becomes progressively less from the northwest to the south--
east so that eastern islands are more arid than those of the west.
2 Distribution and habitat 17
Table 2.1 Distribution, habitat and altitudinal range of the Lesser Sundas Zebra
Finch and other sympatric species of estrildines in the Lesser Sundas archipelago
ranging over 21 islands from Lombok in the west to Sermata in the east (see Figure
2.2)
e I
Lombok
9'-
10'8-
Sawu SemaUa
Sumba ot!!J AÍ Timor
'" Dao"·~Roti Sea
I
115'E
I I
130'E
120'E 125'E
Fig. 2.2 Distribution (stippling) of the Lesser Sundas Zebra Finch Taeniopygia
gutata guttata in eastern Indonesia.
Fig. 2.3 Main distribution of the Australian Zebra Finch Taeniopygia guttata cas-
tanotis. The edge of the stippled area encompasses 507 one-degree blocks in which
Zebra Finches were found breeding or were observed in more than 10% of reports
during the 1977-1981 Field Atlas of Australian Birds (Blakers et al. 1984). This area
constitutes the main distribution of the species although not aH one-degree blocks in
this area reported the presence of Zebra Finches. There were non-breeding records
in an additional 48 one-degree blocks peripheral to the main distribution. Density of
birds was relatively low (report rate <10%) in the peripheral distribution that
mainly extended along the eastern coast of the continent and north into the Top End
of the Northern Territory. Isoclines of two species of estrildines that have extensive
overlapping distributions with the Zebra Finch are shown. Isocline 1 shows the
western and southern limits of the distribution of the Double-barred Finch, which
lives in the north and east of the continent; and isocline 2 shows two zones of distri-
bution of the arid-adapted Painted Finch which lives in the centre and west of the
continent. (Modificd fram Blakers et al. 1984.)
2 Distribution and habitat 19
East Enderby, East Lewis, West Lewis and Dolphin. The widest water
gap is approximately 60 km from the mainland to Barrow island, It is
believed that the island populations are only transitory having being
blown there from nearby mainland sites by high winds during cyclones
and heavy storms (S. J. J. F. Davies and S. Ambrose, pers. comm.)
Zebra Finches are conspicuously absent from Cape York Peninsula of
Queensland and are also absent from a major part of the Kimberley
region of northern Western Australia (Figure 2.3). One can only specu-
late why they have not invaded these regions. They are found in areas
that are harsher and drier than the Kimberley and in fact are abundant at
Wyndham (15°33'S, 128°03'E) and Derby (1 JD18'S, 123°28'E) on the
eastern and southern edges of the Kimberley respectively. Evans and
Brougher (1987) found Zebra Finches at waterholes scattered along a
direct line between Wyndham and Derby, which is further north than the
main distribution shown in Figure 2.3; this extension may be part of the
non-permanent peripheral distribution of the species in this region. The
Kimberley is known as the 'Land of the Finches' for its abundance and
variety of species throughout, yet oddly, the most abundant and
1300 140 0
~~5~-t,'fJi,.
LJarwlIl"
0"
~
!(lMBU?'
~ EY
\f
""-~I
11
·Kathenne
TOP ' (!;
CAPE
YORK
')
OESERT
I
~Allce Spnngs oSandrtngham
\\
-- -~ - --~IE~~!~Nil
I I o
Mlleurao I
:"REA/V!(/orl/'\
I DESUP
j
1:
,
--- -~---
I
I
r -,' t..
Bnsbane o ~
~.
",r (
30°
~
(J
~/ 12'00
130
0
'\
v¿1,Adel",~e~
<"...:::3
1
I '-
'-9r~'thQ -./'1. I oSydney
I Shepparton e -- \
• Melbourne' ->'1
40°--- -- 0_-'0 -_40 0
140° V1 150°
Fig. 2.5 Isohyets of mean annual rainfall in millimetres for arid (:0:;250 mm) and
semiarid (251-500 mm) zones superimposed 011 the mail1 distribution (stippled) of
the Australian Zebra Fi11Ch. (Modified from Leeper 1970.)
extends beyond the 500 mm isohyet to areas of high rainfall towards the
periphery on the east and north of the continent, but high rates of evapo-
ration reduce plant growth here and conditions are dry.
Vegetation associations
Rainfall, in combination with regimes of light and temperature, and
topography are the primary determinants of vegetation patterns in
Australia (Nix 1982). The high rainfall fringes on the eastern and south--
western part of the continent were originally covered in tall, closed
forests, much of which has be en cleared for agriculture since European
settlement. In dry areas farther inland, woodlands dominate in a broad
concentric zone and these gradually give way to low woodlands as it
becomes drier towards the centre; shrublands and grassland dominate in
the areas of the lowest rainfall that form the vast arid interior. The vege-
tation is distinguished by the dominance of the following three genera:
Eucalyptus in the forests and woodlands, Acacia in the shrublands and
Triodia (spinifex) in the grasslands. The height of the vegetation becomes
progressively lower in concentric zones from the moist margins to the
arid centre of the continent. A detailed description of vegetation in
Australia can be found in Bridgewater (1987) and Specht (1981).
Ecology of arid Australia
By comparison with other arid regions of the world the Australian arid
zone, the stronghold of the Zebra Finch, is unusual in climate and ecol-
2 Distribution and habitat 23
ogy. An abundance of perennial plants, especially trees and shrubs, is the
outstanding feature. Although the climate is only moderately arid by
world standards, the unpredictability of climatic extremes over a vast
geographic scale is exceptional (Stafford Smith and Morton 1990). This
is driven by rainfall, which is highly unpredictable in space and time, so
that there is no way to predict the timing and severity of drought. The
occasional big fall of rain, which comes at irregular intervals, is also
important because it structures the landscape so that a mosaic of areas of
varying fertility and moisture is produced. Floodplains and floodouts are
the major areas where nutrients and water are concentrated and here
plant productivity can be higher than the surrounding areas. Neverthe-
less, there are still great extremes in soil moisture and these dominate the
lives of plants and herbivores. In regions with sorne topographic relief,
such as the central ranges, water is redistributed in a more concentrated
and dependable way to so-called 'runon' areas. Here even small falls of
rain can be useful. Therefore, what appears to be a vast, superficially
uniform landscape, is in reality, a mosaic with 'fertile or reliable sites
scattered like islands in a sea of exceptionally infertile and unreliable
conditions' (Stafford Smith and Morton 1990). These islands of higher
plant production vary in scale from crack s in rocks to whole floodplains
and drainage lines, and together form that part of the landscape upon
which Zebra Finches, and many other species of birds and mammals,
rely, especially during drought. These fertile sites provide the grass seeds
and the nesting bus hes for the finches.
Davies (1977a, 1986) maintains that plant productivity in arid regions
is more regular than it appears beca use the redistribution of water to fer-
ti le sites can make even light falls of rain effective; moisture can be sto red
until temperatures in spring and summer permit seed germination. Indeed,
spring breeding on a fairly regular basis appears to be the pattern in most
species of birds in arid southwest Australia (Davies 1979) and might also
be the case in the eastern parts of the arid zone (Serventy 1971).
Habitat requirements of the Australian Zebra Finch
Immelmann (1965b) published a comprehensive analysis of the ecologi-
cal factors that affect the distribution of the Zebra Finch across the con-
tinent with a particular emphasis on those that control the timing and
extent of breeding in different climatic regions. This important article
was published in German and has not been widely cited since publica-
tion. However, it is worth detailed examination particularly in the light
of additional data accumulated over the last 30 years. Detailed ecological
and physiological considerations will be made in Chapters 4 and 5.
Grass and surface water
Daily access to supplies of grass seeds is the essential requirement for the
survival of Zebra Finches and, although they are renowned for their
24 THE ZEBRA FINCH
Rain
Zebra Finches feed almost exclusively on grass seeds (Chapter 4) and can
survive for many months on dry seeds that have fallen to the ground six
or seven months previously. Half-ripe seeds and green leaf material are
2 Distribution and habitat 25
probably necessary for the raising of young (Immelmann 1962a, 1965b;
Chapter 7). The breeding distribution of the species across Australia
depends on the germination, growth and seeding of grasses which, in
turn, depends on the amount and timing of precipitation. Warm, moist
conditions are optimal for the seeding of most species of grasses.
Immelmann (1965br asserts that Zebra Finches are adverse to heavy
sustained rainfall. In comparison with other species of estrildines living
in the Wyndham area of the extreme northern coast of Western
Australia, he found that Zebra Finches were less active during the
continuous downpours of the wet season. They avoided foraging in wet
vegetation and avoided contact with damp lea ves and grass. Although I
could not confirm these observations in central Australia (Alice Springs)
or in northern Victoria, I could confirm Immelmann's observation that
the thinly roofed nests provide no protection against heavy rain, which
can easily penetrate the nesting chamber, and occasionally lead to the
death of nestlings or to the desertion of clutches (Immelmann 1962a).
During the northern wet sea son in 1959, Immelmann (1965b, 1970)
found that Zebra Pinches disappeared from the Wyndham area with the
first heavy falls (November), but returned at the start of April after the
heavy rains had diminished. Breeding was squeezed in during the few
weeks of light scattered showers that preceded and followed the storms
of the main wet season since any nests of Zebra Finches that remained
were knocked to the ground by violent storms and all breeding attempts
failed. Immelmann concluded that this inability to adapt to heavy precip-
itation and wind characteristic of the monsoons of northern Australia is
responsible for movements away from these regions during the wet
season-the heavier the wet season, the further south they move to avoid
the rain. He maintains that across a broad belt of northern Australia,
from Derby (17°18'S, 123°38'E) in the west, to Katherine (14°40'S,
131°42'E) in the east, Zebra Pinches move northward occupying coastal
regions during the dry season (May to October) while at the start of the
wet season they retreat inland again. Consequently, there is a zone along
the north-west and northern coasts of Australia where the Zebra Finch
only occurs during the dry season. Immelmann provided no evidence in
support of this seasonal movement-no banded birds were followed nor
any seasonal changes in distribution reported. During fieldwork for the
Field Atlas of Australian Birds (1977-1981), observations of Zebra
Finches were much lower during the summer months over the whole dis-
tribution, especially in the far north of the continent when few observers
are about. Therefore, there is a bias toward more frequent records dur-
ing the winter months. In spite of this, data from four sites on the north-
ern limits of the range of distribution show that there is a significant
increase in the proportion of Zebra Finches observed during the dry win-
ter months and fewer during the wet summer months (Table 2.2). This
finding is consistent with Immelmann's hypothesis. These movements are
26 THE ZEBRA FINCH
Table 2.2 Occurrence of Australian Zebra Finches on the northern limits of the
main distribution in the wet-dry tropics by month of observation a
aData were extracted from the Field Atlas of Australian Birds, 1977-1981 (RAOU). Numbers show
per cent of observations per month for four northern regions. The main wet sea son extends from
December to April. Greater absolute numbers of observations of al! species were made in the months
of winter (June to August) and spring (September to November) and fewer in autumn (March to
May) and summer (December to February). Ncvertheless, a signif1cantly greater proportion of Zebra
Finches were observed on the northern limits of the distribution during the winter (dry season)
months and fewest in the summer (wet season) months (G = 47 ..51, df = 3, P < 0.0001). This sup-
ports the suggestion that Zebra Finches move away from the northern extremes of the distribution
during the wet season.
not conspicuous; they were not detected in the formal analysis of the
atlas data (Blakers et al. 1984), nor were they confirmed by contributors
to Australian Bird Count (RAOU, unpublished data), a project specifi-
cally aimed at detecting movements by means of changes in seasonal
abundance. It is not surprising that movements are difficult to detect in
the far northern parts of Australia since observers are sparsely dis-
tributed and travel during the wet sea son is difficult because of the inun-
dation. Final confirmatíon of these sea son al movements must wait until
banded birds are retrapped or recovered; however, they are consistent
with Nix's (1976) predictions based on climatic factors.
Immelmann (1965b) also hypothesised that the intensity and timing of
the monsoon in the northeastern part of Australia, in the Cape York
Peninsula region of Queensland, prevented Zebra Finches from breeding
at aH, and so constrained the species from advancing north of latitude
1 rs. He argued that the heavy downpours that began suddenly at the
onset of the wet season made breeding impossible; furthermore, even in
the dry season, occasionally rain was too heavy for breeding. However,
in north Queensland, Zebra Finches normally breed throughout the yeal'
2 Distribution and habitat 27
except for the dry winter months (Chapter 7) and they have been
recorded breeding at Ingham (18°43'S, 146°10'E) in March-one of the
wettest months of the rainy season in one of the wettest regions of
Australia (mean annual rainfall of 2,000 mm; White 1946). Clearly,
Zebra Finches in Queensland are not prevented from breeding by heavy
rain per se; nevertheless, sorne environmental factor(s) prevents them
from breeding north of the 17°S latitude.
Habitat selection
Zebra Finches prefer open grassy country with a scattering of trees and
bushes. Grasses provide food, and trees and bushes provide nesting sites
and shade. Forests, dense woodlands, including mallee, are not pene-
trated by Zebra Finches and are barriers in sorne peripheral parts of the
continent to areas that might otherwise be suitable (Table 2.3). These
formations, which have limited seeding grasses, are also avoided by
Zebra Finches when encountered within the distribution. In the savanna
of northern and northwestern Australia, Zebra Finches also avoid the
dense forests and woodlands that border the banks of rivers and creeks.
It is possible that the dense low tree cover in Cape York Peninsula is one
factor that has prevented colonising.
Zebra Finches have extended their range in the southeastern (Davies
1977b) and southwestern (Immelmann 1965b) margins of the distribu-
tion where forests and woodlands ha ve be en cleared or thinned for agri-
culture and pasture grasses cultivated. Here they are usually found where
suitable nesting trees and bushes are growing. Densities in rural land are
highest in irrigated regions where agriculture is most intense. Grasses
seed in these regions throughout much of the year and this regular
supply of food is shielded from the effects of droughts that prevail
throughout the range. Nesting sites for breeding and roosting are
provided in shelter belts planted around farm houses and outbuildings
and in the numerous fruit trees (especially citrus) and vines cultivated in
the fruit growing regions. It is not uncommon for wild Zebra Finches to
live on the outskirts of towns and in villages, frequently invading parks
and residential areas. I have even seen them breeding in dense shrubs
planted in traffic islands in small shopping centres where they fearlessly
go about their business feeding on cultivated lawns, and raising families
oblivious to people and passing vehicles.
Although Zebra Finches can roost like most birds by clinging onto
twigs or branches they show a strong preference for roosting in nests,
either complete or partially complete. Sites for breeding are needed
especially those that can provide a firm anchorage for the flimsy nests.
Densely branching, preferably thorny, trees and shrubs pro vide the best
sites but other structures, both man- and animal-made will do (Chapter
6). Extensive areas of saltbush and spinifex that are devoid of shrubs and
trees are avoided beca use the required sites for nesting are not present
28 THE ZEBRA FINCH
Table 2.3 Vegetation associations of the Australian mainland within and beyond
the area of the main distribution of Zebra Finches; the quantitative breakdown is
based on the dominant vegetation in 1 blocks a in which Zebra Finches were
0
recorded present or absent during the Field Atlas of Australian Birds from
1977-1984 (Blakers et al. 1984); Zebra Finches are not found in forest, but will
invade margins of forests that have been cleared
a Measures one degree of latitude by one degree of longitud e giving an area of 10,000 km 2 •
b Proportion of 607 block s in which Zebra Finches were recardcd.
( Proportion of blocks of each vegetation formation where Zebra Finches were absent within the
area covered by the main distribution.
Summary
Zebra Finches have an extensive distribution in Australia and Indonesia.
The Lesser Sundas Zebra Finch is found on 18 of the 21 main islands
that make up the archipe1ago. It is found from Sermata in the east to
Lombok in the west, fram sea leve1 to e1evations of 2,000 m. Grasslands,
rice fie1ds, and secondary woodland are the preferred habitats. Clearing
of forests for timber and cultivation has extended the area of suitable
habitat and provided sources of grain. The Lesser Sundas Zebra Finch
has the most extensive distribution of 11 species of estrildine found in
the Lesser Sundas archipe1ago.
The Australian Zebra Finch is found over 75% of the mainland; it is
absent from T asmania, but has reached a number of islands within 60
km of the mainland coast. It is the most wide1y distributed Australian
estrildine; its congenor, the Double-barred Finch, is the next most
widespread species occupying 32 % of the mainland. The Zebra Finch
is not found in Cape York Peninsula or the northern Kimberley, but
occupies all the arid zone of the interior and most of the semiarid zone
towards the periphery of the continent. Dense vegetation and heavy
monsoonal rains limit the extent of its distribution in the north and east,
but clearing of vegetation has extended the range into semiarid peri-
pheral regions. Low temperatures may restrict its spread to cleared areas
in southern parts of the range. Zebra Finch habitats must have accessible
surface water for drinking, grass seeds for food, and bushes and shrubs
for nesting and raosting. In much of arid Australia these resources
are patchily distributed and concentrated most1Y on sites of higher soil
fertility and moisture that occur in flood plains and along water courses.
3 MorphologyJ domestication and moult
Morphological variation
Subspecific variation
Australian Zebra Finches are significantly bigger than Lesser Sundas
Zebra Finches (Clayton et al. 1991). Saroples froro northern Victoria
(Danaher colony) in southeastern Australia were heavier (males 1 g
heavier on average, females 1.5 g) than those trapped on Lombok,
Sumbawa, Flores and Timor; wings and bills were also longer (5 mm and
1 mm on average, respectively). Bill-depth was significantly smaller only
in the Flores and Timor samples (Figure 3.1). Bills of Australian males
were redder, darker and more intensely coloured than those from
the Lesser Sundas, but there were no significant differences between
females. Within each subspecies the bilIs of males are, on average redder,
darker and more intense than those of females which are an orange-red
colour. Considerable overlap exists in bill colour of the sexes (Burley and
Coopersmith 1987). Detailed comparisons of iris colour of the two sub-
species have not been made but both are a deep reddish orange in wild-
caught birds, but a dark brown in aviary-bred ones.
Males of the two subspecies, but not ferr.ales, also differ in plumage on
the throat, fare neck and breast. The fine black barring of the throat and
fore neck found in Australian males is absent from Lesser Sundas males
which simply have a pale grey ground colour. The size of the male black
breast band varies considerably among Zebra Finches from the different
islands of the Lesser Sundas, but in every case the absolute and relative
size of the band is significantly smaller than that found in the Australian
subspecies which is also highly variable. FinalIy, males of the Lesser Sun-
das subspecies have an abdomen off-white in colour, thus resembling
that of females with their honey-coloured wash, whereas Australian
males have apure white abdomen, when free of dirt. In museum skins,
ageing causes the white abdomen of the Australian males to brown
slightly and this tends to make the abdomen resemble that of the
females. Sexual dichromatism is thus more extreme in the Australian
subspecies.
Cross-fostering between subspecies in aviaries does not affect size,
plumage and bill colour differences. Hybrids of both combinations (male
guttata x female castanotis and male castanotis x female guttata) were
indistinguishable in size, with bilI and plumage intermediate between the
two parental subspecies (Clayton 1990a). Interestingly, it was easier to
Males Females
Wing length
E
555
.J:
'é,
c:
..S!
el)
L S F T A L S T A
Weight
L s F T A L s F T A
Billlength
~ 9.5+----------
E
5
-5el)
c:
..S!
¡¡;
L s F T A L s F T A
Bill depth
6.
E
5
-5a.
'"
-o
¡¡; s.
L s F T A L s T A
Fig. 3.1 Size comparison of (A) Australian n = 55 males and 45 females) and
(L,S,F,T) Lesser Sundas Zebra Finches (hatched bars) based on four measurements
of free-living birds. Bars show means plus one standard deviation. L = Lombok (11
males, five females), S = Sumbawa (four males), F = Flores (12 males, 14 females),
T = Timor (eight males and five females). (Modified from Clayton et al. 1991.)
32 THE ZEBRA FINCH
obtain hybrids from the latter pairing than the former (N. Clayton, pers.
comm.).
Table 3.1 Weight (g) of free-living adult Australian Zebra Finches from two dis-
tant regions-northern Victoria (Danaher colony) and central Australia (Alice
Springs Research Centre colony)"
Table 3.2a Means ± s.d. and (range) of four measurements of adult Zebra Finches
from Danaher, northern Victoria
Males Females
a Differences between the sexes: n = 242 males, 237 females F¡477 = 9.6, P<O.OOO1.
b n = 251 males, 241 females F¡49G = 12.0, P = 0.0006. '
'n = 205 males, 205 females F¡'4GB = 9.6, P < 0.002.
d n = 209 males, 287 females F¡',421 = 0.05, P = 0.8.
Table 3.2b Means ± s.d. and (range) of four measurements of adult Zebra Finches
from Alice Springs
24
13
E
o; -S 23
:<:12 -"
O>
'¡¡; -o
~ 22
~
I
11
21
10 20
Ce) (d)
57
E
E E
$ .c:
.c: o,
o. c:
-'"&,
~
c:
-
ro 7 3
Ce)
Fig. 3.2 Size comparisons of four morphometric characters for male and female
Zebra Finches from two widely separated localities, northern Victoria (V) and
central Australia (C). Means (± s.d.) are given. Central Australian birds were signifi-
cantly lighter than birds from northern Victoria (F 1,1627 = 40.1, P<O.OOOl) but larger
in length of head-bill (F¡, 576 = 89.3, P<O.OOOl), tarsus (F I , 518 = 134.0, P<O.OOOl)
and bill-depth (PI, S07 = 1030.0, P<O.OOOl); there was no significant difference in
wing-Iength.
3 Morphology, domestication and moult 35
Domestication
Because it is hardy and easy to keep in captivity the Australian Zebra
Finch has long been one of the most popular cage birds in the world.
There is sorne suggestion that Louis Jean Piere Vieillot held Timor birds
in captivity (Immelmann 1962b, 1965a), but this cannot be confirmed
(Rogers 1986). Captive breeding of the Australian subspecies beca me
well established in Eurape by the 1870s as indicated by the low market
price of the times (Sossinka 1970). Thus, imports of wild stock fram
Australia became superfluous towards the end of the nineteenth century
and gradually diminished, with few reaching Europe after the First
World War (Immelmann 1962c). Export bans were implemented by the
Australian government on all native wildlife in 1960 and no wild Zebra
Finch stock has been legally imported by overseas bird keepers since that
time. By 1962, Immelmann calculated that sorne 50-80 generations of
Zebra Finches had been bred in Europe without any significant input of
wild blood and a fully domesticated strain had arisen.
The first colour morph identified was the 'fawn', which was discovered
in wild birds fram South Australia in 1927. This formed the breeding
stock from which many other mutant strains were developed, and it is
still one of the most popular shown in competitions today. About 30
main colour morphs, or plumage mutations, of domesticated Zebra
Finches current1y exist and there are about 44-50 combinations of these
(Immelmann 1965a; Martin 1985; Rogers 1986; Corbett 1987;
Appendix 1). Homozygous strains are rare. Sorne mutations change the
colour of aH the plumage in both sexes (e.g. white, silver, cream), while
others change particular parts, principally the dorsal surface (e.g. fawn,
penguin), back (light- and saddle-back), breast and thraat (black- and
orange-breasted), cheeks (silver-, grey-, fawn- and black-breasted) and
eye/tear stripe (e.g. black-faced). Relatively few mutations have altered
the colours of the flanks, tail, upper-tail coverts and rump. Sorne muta-
tions are sex-linked (e.g. fawn, chestnut-flanked and saddle-back), and
while most are recessive sorne autosomal mutations are dominant (e.g.
crested, silver and cream). Those strains that change the ground colour
of the plumage (e.g. fawn, silver) produce different combinations with
other strains that change the colour of particular markings. Sorne muta-
tions are based on the absence of pigment, totally (white) or in part
(pied) and the latter can be combined with a number of other mutations.
One autosomal recessive mutation makes the bill colour yellow. Finally,
there is one morph that is not a colour variant, namely 'crested', which
causes a flat swirl of feathers to graw on the forehead. Breeding of these
morphs and the development of new ones are the principal pleasures of
many bird keepers, and Zebra Finch societies have been formed in
Australia, the United Kingdom, Germany, Denmark and elsewhere to
foster this hobby. These societies publish the standards for the various
36 THE ZEBRA FINCH
morphs and the crossings necessary to obtain them and conduct show
competitions.
Immelmann (1962c) and Sossinka (1970, 1972a, 1974, 1980) investi-
gated morphological, physiological and behavioural changes in Zebra
Finches that have arisen as a result of domestication in Zebra Pinch
stocks held in Europe. Offspring of domesticated Zebra Finches were
found to be significantly heavier than aviary-reared offspring of wild-
caught birds. Domesticated females had longer wings, and longer and
deeper bills than their wild counterparts and took longer to reach adult
size, but the males did not differ significantly in these measurements.
These differences arose through isolation of the captive populations,
through chance fixation of genes, and through selection in the captive
environment. Immelmann assumed that most of the ancestral wild-stock
that formed the basis for the domesticated line carne from northwest
Australia, but unfortunately no morphological measurements have been
made on birds from this region. Selection by bird keepers is considered
the strongest force in the domestication of the Zebra Finch (Sossinka
1982).
In contrast to their European counterparts Australian aviculturalists
have not selected for larger body size. Whites, fawns and mixed morphs
were aH smaller than wild-caught Zebra Finches and their first genera~
tion aviary-reared offspring (Carr and Zann 1986). They were lighter,
had shorter wings, head-biHs, tarsi and bill-depths. There is no plausible
explanation for these contrasting directions in the domestication process.
Conceivably, differences may have arisen beca use of difference in sources
of wild stock, in climatic conditions, and in show standards for size.
Morphological differences can arise within one generation of captive
rearing. Colour of the irides is the most noticeable difference: first-gener-
ation aviary-bred birds do not develop the red iris of free-living adults; a
few develop reddish-brown shades, but most retain the dark grey-brown
colour of juveniles (Sossinka 1970). First-generation aviary-bred birds
also have smaller wing-Iengths and shorter tarsi than wild-caught birds
(Carr and Zann 1986). No differences with wild birds are found in bill
colour (Burley et al. 1992), but the grey dorsal surface of the back of
domesticated wild-type males has less of the brownish wash found in the
free-living ones; a consequence of having been selected against by bird
keepers (Sossinka 1970).
Moult
Like aH estrildines Zebra Pinches have nine primaries and nine secon--
daries. The tenth primary is much shorter, being only about one fifth as
long as the resto The more proximal primaries, numbers 1, 2 and 3
(41-42 mm), are shorter than more distal ones (44-47 mm), of which
3 Morphology, domestication and moult 37
numbers 6 and 7 are the longest. All twelve rectrices are the same length
(35-38 mm). The feather tracks (pterylae) for the contour feathers fol-
low the conventional passerine pattern (Ginn and Melville 1983). Only
about 20 down feathers grow from the large ventral apterium. Details of
the pterylography of the Zebra Finch can be found in Parsons (1968).
In sorne males the juvenile body plumage may begin to be replaced by
the secondary sexual plumage as early as 20 days after hatching, but in
most this occurs between days 35 and 40. The first indication that a
juvenile is a male will be a few flecks of colour among the grey plumage,
for example, a black spot on the breast band, an orange spot on the ear
coverts, or odd striped feathers on the throat. Around 40 days, patches
of male-specific feathers begin to appear on the throat and flanks and the
characteristic adult plumage is complete by days 55 to 60 but the
colours, which are somewhat faded at first, intensify in the next few
weeks. For unknown reasons the development of the male plumage is
often delayed for several months in wild birds hatched in late autumn in
south-east Australia.
Acquisition of male plumage marks in the Lesser Sundas Zebra Finch
ha ve not been described.
Wing moult
Wing moult of the Australian Zebra Finch was investigated at the
Padgett colony (Figure 3.3) from 1976 to 1983 (Zann 1985a). Multiple
captures of individuals during the same moult cycle permitted accurate
estimates, through extrapolation, of the starting and completion dates of
primary moult in adults and young. The major finding was thai Zebra
Finches moult their primaries extraordinarily slowly in a continuous
step-wise moult. Feathers are lost in a rigid, non-overlapping sequence in
ascending order from number 1 (the innermost) through to number 9
(the outermost), that is, from proximal to distal. Primaries are not lost
until the preceding one is fully grown, whereas most species of birds will
loose and regrow two or three primaries simultaneously, and so com-
plete the cycle more rapidly. On average, the smaller inner primaries of
Zebra Finches take about 21 days from loss of the old feather to the full
growth of its replacement, and the larger outer ones take about 26 days.
The interval between the full growth of one primary and the loss of its
more distal neighbour was highly variable, ranging from a few days to
several weeks. It was extremely rare for moult to stop during mid-
growth of a primary. Secondaries were also replaced very slowly, but did
not begin before the primary moult commenced: their sequence of
replacement of individual feathers was not so rigid as that of the
primaries, but the approximate order was the conventional one found in
most passerines, namely the tertials first, usually the middle one
(number 8), then the inner (number 9) and then the outer (number 7);
38 THE ZEBRA FINCH
145 0 30'
36°
Numurkah
Wunghnu
• t'.
t* /* ---.'
P a dge
el over lea * t_"J Danaher
* Bunbar t ha
5km
Fig. 3.3 Locations (,:.) of five Zebra Finch colonies studied in northern central
Victoria.
the secondaries then moult in arder from the most distal to the most
proximal (number 1 to number 6).
This slow moulting occurs continuously so that there is only a short
period between successive cycles. Moreover, in many individuals the new
cycle will begin befare the previous cycle is complete, that is primary
number 1 will be lost befare number 9 has been lost or finished growing,
consequentIy there will be two actively moulting centres within the pri-
maries oE each wing. This type oE overlapping moult sequence is termed
'continuous step-wise' moult, 'Staffelmauser' or 'serially descendent'
mouIt (Ginn and Melville 1983). During any month of the year a major-
ity oE Zebra Finches at Padgett were Eound with sorne primaries in pin or
regrowth including those months when breeding was at its peak.
The loss of primary number 1 oE the juvenal mouIt occurred at a mean
age oE 80 days in both males and females, but the latter too k significantly
longer to complete the cycle (males 204 :!: 54 days vs. Eemales 224 ± 56
3 Morphology, domestication and moult 39
days). The date of hatching had a significant effect on the age when juve-
nal moult began. y oung hatched in the spring half of the breeding season
commenced their moult before 80 days of age, sorne as early as 30 days.
Those that hatched in the second half of the season waited until they
were 80-100 days of age, and those that hatched in April, the last month
of the breeding season, waited until winter was over before they lost
their first primary. In many young the next cycle of primary moult began
before primary number 9 of the juvenal moult was complete so that the
two cycles overlapped.
The duration of primary moult in adults at Padgett was slightly longer
than that of the juvenal moult, but there were no significant differences
between the sexes (males 229 ± 62 days vs. females 239 ± 79). The
interval between successive moults was longer in females (mean 6 days)
than males (mean -16 days), that is, their successive moult cycles over-
lapped less; consequently, a greater proportion of males than females
captured in the course of the study had primaries in moult (males in
moult 68% VS. females 58%, n = 3,080). There were no significant dif-
ferences in duration between successive cycles, either from juvenal to
first, or from first to second. Primary moult in adults could be initiated in
any month of the year, but rate of feather growth and replacement was
slowest in winter and early spring, presumably beca use of low tempera-
tures and food shortages. The amount of nesting activity each month
was inversely correlated with the rate of primary moult, which suggests
that the two processes are competing for the same sources of energy
(Zann 1985a). This appears to be more so in females than males since
significantly more males were found in moult than females during the
three peaks of the breeding sea son at this colony (Chapter 4). Neverthe-
less, individuals could breed and moult simultaneously, as 1 found when
1 raided breeding pairs in their nests. Of 12 parents caught one night in
March 1981, seven (four males and three females) were in active stages
of feather re-growth and one female was growing two primaries (num-
bers 9 and 1) simultaneously.
Zebra Finches at other colonies in northern Victoria (Shepparton,
Cloverlea and Danaher; Figure 3.3) had a similar sequence and duration
of primary moult to that found at the Padgett colony. Furthermore,
Zebra Finches at Alice Springs demonstrated a basically similar pattern
of moult to that found in northern Victoria (R. Zann and N. Burley,
unpublished observations); however, a smaller proportion of trapped
adults were found in active moult at any one time (28% at Alice Springs
vs. 63% in northern Victoria) and more had moult stopped mid-cycle so
there were fully grown primaries of two or more distinct ages (based on
colour and degree of wear). Complications often arose when several
simultaneous cycles were stopped and re-started several times on the one
wing. Moult usually restarted at the next distal primary in both cycles
('suspended moult'), but in sorne instances it also initiated a new cycle
40 THE ZEBRA FINCH
mouit activity and its step-wise progression among the primaries became
increasingly obscure.
This study of wing mouit in wild-caught and aviary-bred Zebra
Finches is instructive, for three reasons. First, it confirms all the main
findings in free-living birds; second, it demonstrates the flexibility of the
moulting processes in response to environmentai conditions, especial1y
the availability of food and the incidence of breeding, and third, it rein-
forces the need for caution in interpreting data obtained from captive
subjects.
Summary
Lesser Sundas Zebra Finches are about 13-18% smaller than Australian
Zebra Finches. Plumage of the females is identical, but Australian males
have redder bills and a proportionally larger black breast band than
Lesser Sundas males. Sexual dimorphism is more extreme in the Aus·,
tralian subspecies. Females are heavier than males in breeding and non-
breeding seasons, but males have longer wings, and longer and deeper
bills; there is no sex difference in tarsus length. Zebra Finches from cen-
tral Australia were larger (head-bill, bill-depth, tarsus), but not as heavy
as those from southeastern Australia. The harsher climate in central Aus'·
tralia may be responsible for the lower weight and poorer physicai con·'
dition. Biochemical investigation of the Australian subspecies may reveal
sorne degree of geographic differentiation.
Captive Zebra Finches have become highly domesticated and were first
bred in Europe almost 130 years ago. There are about 30 main strains of
colour morphs of the Australian subspecies in captivity and about 50 dis,-
tinct combinations of these. In Europe, domesticated females are larger
and heavier than offspring of wild-caught birds and males are heavier,
although they are not larger. By contrast, in Australia, domesticated
colour morphs are smaller and not as heavy as wild-caught birds and
their first-generation offspring. Different selection criteria among bird
keepers may be responsible for these differences. First-generation aviary-
bred birds do not develop the red iris of free-living birds.
In free-living birds the secondary sexual plumage begins to replace the
juvenile plumage around 35-40 days after hatching, and is complete by
about 60-70 days and reaches the full coiour intensity of adults by
90-100 days. The rate of development is not uniform and is slowest for
young hatched in the autumn months. Free-living adults moult the wing
feathers in a very slow, continuous step-wise manner and the whole cycle
takes about 235-290 days. The time from 10ss to full growth of the
replacement primary takes about three weeks. The next moult cycle may
start before the previous cycle has finished, so that most birds are i.n a
state of permanent primary mouit, although a greater proportion of
3 Morphology, domestication and moult 43
birds in central Australia stopped their moult during a cycle than did
birds from northern Victoria. Young begin their first primary moult at a
mean age of 80 days after hatching but the actual age depends on when
young hatch. Adults may moult during a breeding cycle, but do so more
rapidly outside the breeding season. Females moult more slowly than
males. In captivity intense breeding activity slows wing moult and the
order of feather replacement is gradually obscured. In unfavourable cli-
mates and during harsh seasons moult slows and may frequently stop
until conditions improve. Among passerines only the Double-barred
Finch, the Long-tailed Finch and the Masked Finch moult as slowly as
the Zebra Finch. The main advantage of slow moult is its minima"!
impairment of flying ability so that birds maintain maximum mobility,
an essential requirement for survival in arid and semiarid habitats for
sorne speCles.
4 Feeding ecology
"It passes much of its time on the ground, and feeds upon the seeds of various
kinds of grasses.'
J. Gould 1865.
Diet
Methods
Three methods have been used to investigate the diet of Zebra Finches.
Immelmann (1962a, 1965a) simply observed what birds were feeding on
and collected a sample for identification. This technique is difficult with
fallen seed beca use often nothing is found when the ground is examined.
Zebra Finches have been shot in two studies (Davies 1977a; Morton and
Davies 1983) and the contents of the crops identified. Identification can
be a problem since the characteristics of the dehusked seed are different
from those of whole seed. Consequently, uncertainties arise, especially
when several congeneric species have been eaten, and extensive reference
collections of dehusked seeds are necessary. The virtue of shooting is that
it is simple and efficient, and the whole contents of the crop can be
examined, but it is a bit drastic for the unfortunate bird. The third
method is a non-destructive one devised by Zann and Straw (1984b): a
sample of seed from the crop of mistnetted birds is extracted by means of
a narrow plastic tube. This tube is gently pushed down the oesophagus
into the crop's store of seed where individual seeds are pushed up into
the bore of the tube by manipulating the wall of the crop with a finger.
4 Feedi1zg eeology 45
Up to ten to fifteen seeds can be extracted and if the crop contains water,
capillary action simply draws seed and water up. Schopfer (1989)
bevelled the tip of the tube to allow large seeds to entero The tu be
method is not very aversive for the birds and it provides representative
samples of the types of seeds in the crop, although it underestimates the
abundance of insects (Zann and Straw 1984b). Where the diet is fairly
well known, conspicuous types of seeds can be identified through the
thin, almost transparent wall of the crop (Tidemann 1987), but samples
are required for confirmation and reference.
seeds. Insects were extremely rare in the crops of 510 Zebra Finches shot
by Davies (1977a) in arid western Australia. Immelmann (1962a, 1970)
saw Zebra Finches taking flying ants and termites in mid-air in northern
Australia. Birds would leave a lookout perch and make short fluttering
hawking flights then swallow the prey on landing again. Remains of
snails and ostracods ha ve also been found in sorne crops but these were
probably eaten for their mineral content rather than as food. Green
material, often lea ves and seed heads nipped from the extreme tips of
grasses and succulents, is often found in the crops during the breeding
sea son in both adults and nestlings (Zann and Straw 1984a). Fine gravel
and particles of sand are also frequently found in crops.
No significant sex differences in diet were found in any of the eight
quantitative studies (Table 4.1).
Fowlers Gap, N. S. W. 45 5 12 Enneapogon sp. 1.67 ± 0.47 Morton and Davies (1983)
Sandringham, Qld. 52 5 14 Panicum decompositum 1.66 ± 1.02 Morton and Davies (1983)
Northern Victoria 413 15 12 Echinochloa crus-galli 1.41±0.44 Zann and Straw (1984)
Burrendong, N. S. W. 14 1 7 Danthonia sp. 1.32 Schopfer (1989)
Undoolya Bore, N. T.t 39 1 4 Enneapogon cyclindricus 1.06 Zann et al. (1995)
Alice Springs, N. T.* 23 1 2 Cenchrus ciliaris 1.05 Zann et al. (1995)
Alice Springs, N. T.§ 25 1 5 Cenchrus ciliaris 1.19 Zann et al. (1995)
Ooraminna Rock Hole, N.T.~ 18 1 7 Digitaria brownii 3.15 Zann et al. (1995)
n 2
., B = 1 / LP¡ where p¡is the proportion in the diet of the i th seed type in the sample and n the number of seed types (Wiens and Rotenberry 1979).
i=!
t 31 km E of Alice Springs (23°45'S, 134°06'E).
* Settlement Tanks, 6 km SW of Alice Springs (23°44'S, 133°50'E).
§ Airport, 13 km S of Alice Springs (23°48'S, 133°52'E).
~ 30 km S of Alice Springs (24°02'S, 134°00'E).
48 THE ZEBRA FINCH
c°,-~--------------------------------------------------,
~ JJ
=>30 50 ~"
~ :J
~
Q)
o..
§20 40
¿
"3
"E 2-
-0 10 30
e
'"X .,
~ O f----
20
Danthonia caespitosa
::::::
Irrigation season
Fig. 4.1 Schematic portrayal of seasonal growth and production of five species of
grass seeds heavily utilised by Zebra Finches in an irrigated regio n of northern
Victoria. The top panel shows mean monthly rainfall and ranges of mean monthly
maximum and minimum temperatures. (From Zann and Straw 1984a.)
Narrow diet
Despite the large range of sizes and species of seeds capable of being
eaten by Zebra Finches, quantitative studies of diet at eight locations
showed that they concentrated on seeds from a limited number of grass
species. For example, in northern Victoria 66% of seeds eaten over a
15-month period were E. crus-galli (Zann and Straw 1984a) and in the
eastern arid zone Morton and Davies (1983) found that 84% of seed
eaten at Fowlers Gap (New South Wales) was an Enneapogon species
and 74% eaten at Sandringham (Queensland) was Panicum decomposi-
tumo Mean dietary diversity B, which is an index of the breadth of the
diet, ranged from 1.05 to 3.15 (Table 4.1). Indices for northern Victoria,
Sandringham and Fowlers Gap were not significantly different (Zann
4 Feeding ecology 49
and Straw 1984a) and are similar to that found at Burrendong (New
South Wales) by Schopfer (1989). The greatest diversity was found for
the Ooraminna population near Alice Springs where B = 3.15. This pop-
ulation was in poor condition and few seeds were found in the crops, but
they carne from seven species (R. Zann and N. Burley, unpublished
observations). Tidemann (1987) studied the diet of Zebra Finches, Goul-
dian Finches and Long-tailed Finches in the Top End of the Northern
Territory. Unfortunately, her index of diversity was different from that
used here so comparisons are difficult. Zebra Finches in her study area
had the most di verse diet, eating the seeds of 11 species of grasses,
whereas the Gouldian Finch ate only three species. Zebra Pinches had
the least dietary overlap when compared with the other species.
Morton and Davies (1983) were impressed by the dietary specialisa-
tion of the Zebra Finch populations they investigated, although they did
not assess the proportions of seed available to determine if thediet
departed from random. They compared the breadth of diet to that
reported in the literature for a number of other species from different
parts of the arid zone. Breadth of diet was not significantly different
from that of Budgerigars (Wyndham 1980) which ate a smaller propor-
tion of grass seeds than Zebra Pinches, but breadth of diet was narrower
than that of the Spinifex Pigeon Petrophassa plumifera which also ate
seeds from many species of shrubs (Morton and Davies 1983). Larger
granivores eat a more di verse diet than smaller granivores, presumably
because larger bills enable them to eat a larger size range of seeds. The
same conclusion has been reached by many others who have investigated
the diet of avian granivores (e.g. Newton 1967; Willson 1971; Abbott et
al. 1977).
Zebra Finches had a significantly narrower diet (B = 1.32) than that of
four other species of finches studied by Schopfer (1989) at Burrendong
(Red-browed Finch, B = 3.57; Diamond FiretaiI, 2.66; Plum-headed
Finch, 2.27; Double-barred Finch, 2.67). However, Zebra Finches were
uncommon at this study site and all the samples (n = 14) was taken in
the same month, hence the sample may not be representative and the diet
not as narrow as thought. What was clear from Schopfer's study was
that Zebra Finches take fewer non-seed items than the other species of
finches and do not increase their intake of insects during the breeding
season.
abundant in the arid zone and provide a relatively stable and accessible
resource in this environment. Perennial grasslands cover 40% of the arid
zone and, in addition, form an important understory layer in the Acacia
and Eucalyptus shrublands that form another 41 % of the arid zone.
These shrublands with their grassy understory also ext~nd into many
parts of the semiarid regions and thus allow granivores, such as the
Zebra Finch, to occupy a wide range of country of differing aridity.
Grass seeds are a surprisingly reliable source of food despite erra tic rain-
faH in much of the arid zone. Davies (1977a, 1979, 1986) has shown
that Zebra Finch populations at Mileura (Western Australia) manage to
survive, even in poor seasons, by concentrating on limited parts of the
landscape where grass growth and seeding is most likely. These micro-
habitats normally have sorne annual grasses germinating and seeding
every year and allow sorne Zebra Finches to survive, but not necessarily
reproduce. In good seasons, germinatíon and seeding will be prolific and
may also occur in the flood plains away from the watercourses and allow
Zebra Finches to breed freely. Zebra Finches appear to specialise on
those species of grasses that are adapted to such microhabitats. Not SUf-
prisingly, rainfall is patchy and difficult to predict, so when it fails com-
pletely in one area, Zebra Finches become vulnerable to starvation and
seek seeds elsewhere. Such mobility is essential to survival in arid
reglOns.
Grass seeds, in comparison to seeds of other species of plants, are a
'convenience' food for Zebra Finches. Most grass seeds are a convenient
size for easy dehusking, they do not have hard protective or sticky coats,
and they lack toxic chemicals (Morton and Davies 1983). Moreover,
grass seeds may be produced in great abundance and can persist on the
ground in good condition for many months, especially in a dry environ··
ment, and form a 'seed bank' of food that can be exploited when needed.
The high (80%) carbohydrate content of grass seeds makes them an ideal
food for arid-adapted species beca use they not only provide energy, but
metabolic water as well (Chapter 5). Although granivores constitute only
17% of avian species in the arid zone they make up 44% of all individu··
als and are the most abundant group. However, in completely waterless
regions, such as the Great Victoria desert they are almost absent (Fisher
et al. 1972).
Foraging
Choice of seed
In northern Victoria, Zann and Straw (1984a) found that Zebra Finches
preferred larger seeds over smaller seeds. Echinochloa crus-galli, the
largest seed available, was preferred even when it was old, over small
good-quality seeds, such as Chloris truncata. There was a significant pos-
4 Feeding ecology 51
itive correlation between the percentage of each species of seed eaten by
the finches and weight of individual seeds. Laboratory tests on wild-
caught finches confirmed this preference.
When domesticated Zebra Pinches were raised exclusively on one type
of seed, they showed a clear preference for that type when given a choice
(Rabinowitch 1969). Similarly, in rigorous experiments Palmeros (1983)
showed that domesticated adult Zebra Finches have an initial bias
towards that type of seed with which they were most familiar. However,
when forced to sample other types of seeds and then offered a choice of
four seed types (White Millet Panicum milateum, Yellow Millet Setaria
italica, Japanese Millet E. crus-galli, and Canary Seed Phalaris canarien-
sis) in equal abundance and accessibility, they preferred the type that
provided the highest net gain of energy per unit of foraging time.
'Profitability' was a function of the size of the seed and the ease it could
be manipulated or dehusked, that is, minimum handling time. White
Millet was preferred to Yellow Millet despite taking longer to dehusk
and to swallow; yet its larger size (before husking 3 x 2 mm vs. 2.5 x 1.5
mm) allowed more grams of seed to be ingested per minute; conse-
quently, the daily energy requirement was attained in a shorter time.
Canary seed was the largest seed of all but it was difficult to dehusk.
Costs and benefits of foraging choices in terms of rate of food intake is a
widespread phenomenon in animals (Krebs and Davies 1993).
In their choice of seeds, laboratory Zebra Finches do not always opti-
mise rate of intake of energy. In an early experiment on seed choice,
Morris (1955) offered Zebra Finches White Millet, Yellow Millet and
Canary Seed for two weeks, and in each case found a much greater pro-
portion by weight of Yellow Millet was eaten (56-68% Yellow Millet,
21-41 % White Millet and 2-4% Canary Seed) despite the fact that the
Yellow Millet was smaller and harder to dehusk than the other two. This
more natural and longer experiment may account for discrepancies with
the findings of Palmeros who tested his birds for a maximum of three
days and deprived them for several hours before the experiments began.
Feeding methods
Zebra Finches take seeds individually, mostly ripened ones that have
fallen to the ground, but seed on the heads of standing grass is also
eaten. Seed heads lying on the ground or within reach are pecked out
individually but higher ones are reached by flying up and pecking out
seeds one at a time or by perching on branches at the same level as the
grass head. Birds may jump up, seize the head in the bill or occasionally
grasped it with the feet, and pull it to the ground where it is clamped
with a foot until all the seeds are picked off. 1 have seen birds fly at old
grass heads, take a bill-full and land nearby and dehusk three or four
seeds before flying up to grab another loto Sorne birds, not necessarily
52 THE ZEBRA FINCH
a e e e 40
300
I
, •,,
n
'
30
? • I
, ,,
,,
200 n
,,A--i ,, Z
.,
Q) Q)
N l'
-w ,' ,, Ii'
, 20 ~
,,,
ro
~
,,
\
,,
u
o
¡¡: 100 , I
,
,,, ,,
'll 10 ~,
O O
F M A M J J A S O N D J F M A M J J
1981 1982
Fig. 4.2 AnnuaI changes in the mean size of feeding flocks and the number of
breeding nests by month at the Cloverlea colony, northern Victoria. No samples
were made in June 1982. Birds were not fed at this colony. Error bars are not shown
but were very smaIl during the nesting period, but quite Iarge outside it. Breeding
nests had either eggs or young. The main flushes of ripening seed eaten by the
finches are indicated by Iettering: a, Cynodon dactylon; b, Poa annua; c, Danthonia
caespitosa; d, Amphibromus neesi; e, Echinochloa crus-galli. (Adapted from Zann
and Straw 1984a.)
56 THE ZEBRA FINCH
Feeding flocks
The size and behaviour of feeding flocks are a function of the abundance
and distribution of seed and the level of breeding activity. Feeding flocks
during the breeding sea son were much smaller than during the non-
breeding sea son (Figure 4.2). This is partly due to the way breeding
activity disrupts the freedom to flock, especially in the morning when
incubatíon and provisioning duties are greatest. There was a significant
negative correlation between mean monthly flock size and the number oí
active nests at the Cloverlea colony in northern Victoria (Zann and
Straw 1984a). Mean size of feeding flocks was below 10 in November,
December and ]anuary, but between 10 and 20 during other months of
the breeding season. The large flocks in early spring may include non-
breeding adults and those in the second half of the season may include
young. Once breeding cea sed in April (in 1981) or May (in 1982) there
was a drama tic increase in the size of feeding flocks; mean monthly flock
sizes ranged from 150-350 birds. These large flocks formed in two ways:
a smallish feeding flock would be joined on the ground by other small
flocks, or birds would assemble in dribs and drabs in a nearby tree until
a large flock was formed whereupon they would descend to the ground
and feed as a single flock. There was a significant negative correlatíon
between mean monthly minimum temperatures and flock size during the
winter months but none during the summer months.
The flight-distance of feeding flocks varied seasonally at Cloverlea.
During the breeding season, small flocks could be approached to within
10m before they fled, whereas large flocks could not be approached
within 15-20 m; sorne fled when 40 m away. Larger flocks provide an
increased level of vigilance over smaller flocks but experiments are
needed to determine the size beyond which vigilance no longer increases.
Selfish herd benefits would increase with increased flock size.
Competition for food may limit flock size and its dispersion on the
4 Feeding ecology 57
ground. In winter, E. crus-galli was abundant in large, dense patches at
Cloverlea and flocks were at their maximum, whereas smaller flocks
existed during the breeding season when food was more scattered and in
smaller patches. A similar thing was observed at Alice Springs during a
prolonged dry spell in 1986. Flocks were small in most places as food
was scarce although large flocks could form in areas distant from the
roosting nests. Here they foraged in a coordinated way although they
were wide1y dispersed over the ground (see be1ow), except at one loca-
tion, where irrigation had produced dense patches of Buffe1 Grass seed.
Here, tight flocks of 60-160 would form, and once I caught 40 birds out
a foraging flock of 80 in a single collision with a mistnet. Flock size is
thus a function of both avoidance of predators and avoidance of forag-
ing competition. It is interesting to note that large flocks still formed dur-
ing August at Cloverlea by which time the bank of E. crus-galli seed was
exhausted, hence competition must have be en high. Perhaps large flocks
were still an advantage beca use they enhanced the probability that sorne
large patches of yet unexploited seed may be discovered; alternative1y,
even if birds were starving at this time they could still reduce their risk of
predation by joining a large flock.
When feeding, Zebra Finch flocks are remarkably silent for what is
normally a noisy bird and can cause one to startle when a flock is dis-
turbed in long grass. They must be conspicuous from the air since small
parties flying high overhead will suddenly change direction and dive
down to join a feeding flock. Perhaps the white rump is a re1easer. Flocks
feeding on seeds in standing grass are conspicuous as birds jump up and
pull down heads. A flock feeding on fallen seed move across the ground
in one general direction, the speed depending on the density of seed.
When high, the movement is smooth and gradual with the birds in front
advancing as those behind catch up, but when seed is scarce the move-
ment is fast, jerky and erra tic, with individuals towards the rear taking
to the air and landing in front of the leading birds. This 'roller feeding'
goes in waves with intervals of 20 to 60 seconds.
An undisturbed feeding flock disperse gradually into smaller flocks
from two to ten individual s that commute to the nearest trees or back to
the nesting colony, or off in search of another patch of seed. If suddenly
disturbed, the whole flock instantaneously takes to the air in a deafening
din and rushes to the nearest tree or fence where they alight. First
impressions suggest chaos, but the flock is highly cohesive and well coor-
dinated with pairs and family groups keeping together so that there is lit-
de re-positioning of members of families and pairs after landing.
ing, dry season. These were genuine flocks as distinct from the aggrega-·
tion of species around water-holes and rich patches of seeds. Members of
these flocks maintained their coherence and attraction for one another
beyond localised resources. He described three types of mixed finch
flocks: those whose members aH had red bills, those with grey bills and
those with yellow bills. He believed that the common bill colour was the
optical relea ser that permitted them to flock together. The red-billed
flock consisted of the Zebra Finch and the Star Finch; the yellow-billed
flock, the Long-tailed Finch and the Masked Finch; and the grey-billed flock
consisted of three species of mannikins: the Chestnut-breasted Mannikin,
the Yellow-rumped Mannikin and the Pictorella Mannikin. Strangely,
S. Tidemann (pers. comm.), who for a number of years studied the
Gouldian Finch in the same region where Immelmann made his observa-
tions thirty years earlier, failed to observe any mixed-species flocks of
finches, but only aggregations around water holes. Evans et al. (1985)
observed Long-tailed and Masked Finches forming mixed-species flocks
in the vicinity of waterholes in the Kimberley, but recorded that the
other species noted by Immelmann (above) maintained single-species
flocks. In eastern Australia it is not uncommon for Zebra Finches to be
seen feeding on dense patches of ripening grass with Double-barred
Finches, yet when alarmed the latter fly to heavier vegetatíon while the
former stay out in the open.
At the end of May 1986, 1 observed a mixed-species flock in the
rugged Macdonnell Ranges near Alice Springs. When 40 m up a rocky
slope in Heavitree Gap 1 heard distance calls of Zebra Finches and saw
about 50 birds flitting among the rocks and spinifex; approximately half
turned out to be Painted Finches. These were darker and more upright
than the Zebra Finches and gave a 'tidit-tidit' call every time they flew a
few metres. The Zebra Finches were more conspicuous; they perched on
tops of rocks and called more frequently. The flock was widely scattered
but appeared well coordinated and moved slowly down the slope for
more than 50 m feeding as they went. An Australian Magpie Gym-·
norhina tibicen suddenly landed among the flock and they immediately
took to the air in a single dense formation and flew 30 m to a dead tree.
Sorne Painted Finches then flew off in their own small flocks but the
remainder stayed and gradually moved away with the Zebra Finches,
feeding as they went. Clearly, this was a genuine mixed-species flock,
and interestingly, both species had red bills. Members of both species
appeared to benefit from the relationship in terms of finding food and
avoiding predators, although the following week at the same location 1
saw a similar number of finches of both species, but this time they were
in separate flocks.
4 Feeding ecology 59
Juvenile feeding flocks
Small all-juvenile flocks were commonly formed about a week after
fledging, but were fairly inconspicuous at my colonies in northern Victo-
ria and Alice Springs. Sometimes young from several families formed
creches in thick trees until called out for feeding by returning parents,
but in most cases families kept separate. By the time young are between
30 to 35 days they can usually fly well enough to join their parents in
feeding flocks. y oung less than 30 days of age are often seen feeding
with other colony members up to 400 m from the nearest nesting trees
and on only two occasions have I seen all-juvenile feeding flocks away
from the nesting sites.
Summary
Zebra Finches are grass-seed specialists and unlike most other estrildines
rarely take insects, even when raising young. Seeds from 61 species of
grasses are eaten; nine are exotic species, but Zebra Finches have not
become crop pests. Seeds of a great range of sizes are eaten and each seed
is dehusked before swallowing. There are no sex differences in dieto Food
is limiting during drought in central Australia and at the onset of new
growing seasons elsewhere. Dietary diversity is quite low with birds in
most areas specialising on certain species of seeds; in northern Australia
the diet was more diverse than that of other species of estrildines. In the
arid zone grass seeds provide a rich and fairly dependable so urce of food
for Zebra Finches. Size, familiarity and dehusking time are the important
factors in choice of seed type, but nutrient content is also important.
Most seed is taken from the ground, but half-ripe seed is taken from the
heads of standing grasses. In captivity the daily requirement per individ-
ual is about 3 g of seed while up to 5.3 g is needed in the wild. Search
time and seed abundance determine choice of seed patch in the wild. Net
energy intake of domesticated Zebra Finches during foraging affects
lifetime reproductive success. Half-ripe grass seeds are an important
component of the diet of breeding birds. Outside the breeding sea son
feeding flocks of up to 350 individuals may form but flocks of 10-20 are
frequent in the breeding season. Mixed-species feeding flocks and all-
juvenile feeding flocks are rareo
5 Drinking~water relations and
temperature regulation
Although Zebra Finches have beco me one of the symbols of the Aus~
tralian arid zone, they are not considered among the true desert special-
ists that have evolved in the deserts and can live nowhere else (Frith
1973; Schodde 1982). Zebra Finches are among a number of species,
including many granivores, that have more widespread distributions and
ha ve penetrated the deserts from surrounding semiarid regions. Selection
in these species has tended to favour flexibility rather than specialisation,
and this strategy is exemplified in the adaptations and pre-adaptations of
Zebra Finches to the three main elements of desert environments-·-dry··
ness, heat and cold.
Drinking
Tip-down drinking
Immelmann (1962a, 1966) observed that Zebra Finches swallow with
the bill tip immersed in the water (Figure 5.1), while most other species
of birds bring the bill tip up in order to swallow. The functional mor-
phology of 'tip-down' drinking in the Zebra Finch and 'tip~up' drinking
in the Bengalese Finch Lonchura striata has been studied in detail by
Heidweiller and Zweers (1990) using film and radiogram analysis and
their interesting results are briefly summarised below.
During each bout of drinking the bill is immersed in water for an aver-
age duration of one second and pulsating throat movements occur for
about half a second later as the head is raised. Drinking actions of the
Bengalese Finch are superficially similar except that the bill is only
immersed for about one third of a second. Essential differences between
the species occur during the immersion phase. In the Zebra Pinch, the
fleshy tongue with its arrow head shape is used to scoop up, then pull
back adose of water into the pharynx. The front of the larynx then
immediately forces this dose into the oesophagus where peristalsis trans~
ports the water to the crop. This 'double~scooping' or pulling action
5 Drinking, water relations and temperature regulation 61
Fig. 5.1 A female Zebra Finch uses the tip-down method of drinking to draw up
water from a high position.
occurs whiIe the bill tip is still immersed in the water and the tongue
scoops in and out in very rapidIy (about 50 ms per cycle) so that 18 to
20 double scoops occur in one immersion of the bill. The BengaIese
Finch also uses the tongue to scoop up water into the mouth and phar-
ynx in the same way as the Zebra Finch, but can only do about five
scoops before the pharynx is full, then it must withdraw the bill from the
water and raise it so that gravity and the scooping action of the Iarynx
transport the dos e of water down into the oesophagus. Thus, both
species ha ve a two-scoop mechanism but the Zebra Finch integrates both
into a continuous, uninterrupted sequence whereas the BengaIese Finch
has two disjunct steps. Water enters the mouth of the Zebra Finch twice
as fast as that of the BengaIese Finch and because the Iatter must raise
the bill tip in order to swallow the former can get water into the crop
about ten times faster. Heideweiller and Zweers (1990) found that in the
Iaboratory, both species drank 1.5 mI per day but the Zebra Finch only
needed a mean of 3.6 seconds whiIe the BengaIese Finch needed 39.6 sec-
onds to imbibe the daiIy requirement. Finally, both species could scoop
up very small (0.01 mI) droplets of water equally proficiently. This thor-
ough study also concluded that Zebra Finches do not suck up water by
means of peristaltic movements of the oesophagus, as suggested by
Immelmann and Immelmann (1967), nor do they drink in the same way
62 THE ZEBRA FINCH
Daily drinking
Continuous observation of waterholes has established that most Zebra
Finches come to drink fairly regularly throughout the day if water is
close by (Fisher et al. 1972). When water is beyond about 5 km sorne
observers have detected an increase in frequency of drinking in the hours
around noon (Davies 1971; Evans et al. 1985; Schleucher 1993); this is
the pattern of drinking for all other Australian estrildines, except for the
Poephila which drink once in the early morning and once in the later
afternoon (Immelmann 1965a). Sorne arid zone species avoid drinking
around mid-day, presumably to avoid the high temperatures and intense
solar radiation at this time, but these factors do not deter Zebra Finches.
Almost all keen observers of wildlife in the arid zone have remarked
on the tightness of the link between the availability of surface water and
the presence of Zebra Finches. Zebra Finches are usually found at water-
ing points beca use daily drinking appears essential to their survival and
large numbers congrega te around permanent sources during dry spells.
For example, Fisher et al. (1972) recorded 17,750 drinking visits in a
single day by Zebra Finches at one waterhole north of Alice Springs in
September 1967; this number of visits is probably close to the actual
number of individuals coming into drink since temperatures did not
exceed 23°C on the day the observations were made, so repeated visits
by the same individuals are unlikely.
5 Drinking, water relations and temperature regulation 63
Zebra Finches have been used by thirsty humans to find water
(Serventy 1971). The direction flocks take towards water can be
followed and small, hidden rock holes can be located by listening for
their distance calls. Immelmann (1962a, 1965a) and Fisher et al. (1972)
maintain that Zebra Pinches come to water hourly under summer condi-
tions. This may occur when water is locally available to a breeding
colony or feeding flock, but often water is sorne distance away so that
frequent visits are impossible. When 1 mistnetted for whole days in sum-
mer around bores and rock-holes in the Alice Springs district in 1986 1
never caught a same-day re trap from 250 captures, nor did 1 see banded
birds return a second time to drink. Furthermore, 1 have observed hun-
dreds of Zebra Finches coming to drink each day at bores and dams, but
when 1 searched widely in the surrounding country 1 found few signs of
them. Hence, they probably travel considerable distan ces to water during
dry periods, but it is difficult to establish precisely how far they travel.
Birds may be observed far away from the nearest known sources of
water, but in rocky country it is possible that sorne minute water hole
is hidden away and unknown to all but the finches. However, in
sand country where rock holes do not occur, all bores and wells are
well known to the local inhabitants so that sorne degree of certainty
can be placed on distances birds must travel to drink. Thus, in
the Tanami Desert, sorne 500 km northwest of Alice Springs, S.
Morton (pers. comm.) has seen Zebra Finches roosting at least 25 km
from the nearest so urce of water. Similarly, in the Simpson Desert, Bad-
man (1979) has seen Zebra Finches 27 km from the nearest surface
water at Purni Bore (26°17'S, 136°06'E). Finally, S. J. J. F. Davies (pers.
comm.) observed Zebra Finches living in parts of the Sandy and Gibson
Deserts where there was no water, which lead him to condude that, pro-
viding they can find seed, they do not require daily supplies of drinking
water.
Regular surveys of birds conducted by P. Taylor (pers. comm.) in the
Simpson Desert around Cowarie Station (27°45'S, 138°15'E) in South
Australia provide interesting evidence on the ability of Zebra Finches to
move away from sources of surface water. Zebra Finches were found in
most survey blocks (17 km x 17 km), but in five there was no surface
water and the nearest sources were 12-25 km away. At sorne sites Taylor
observed Zebra Pinches 50-80 km from known surface water. The source
of water in a number of blocks was not surface water but leaves of a suc-
culent, the Large Pigweed Portulaca intraterranea, which is a ground cover
species and is common in sorne areas. Zebra Finches were seen eating
these leaves, as were five other species of granivores, namely the Peaceful
Dove Geopelia placida, the Diamond Dove Geopelia cuneata, the Crested
Pigeon Ocyphaps lophotes, the Galah Cacatua roseicapilla and the Litde
Corella Cacatua sanguinea. Presumably, these birds extracted enough
preformed moisture to forego daily intakes of surface water. Eco-physio-
64 THE ZEBRA FINCH
Water relations
Daily requirements
While there are no studies of the amount of drinking water needed by
free-living Zebra Finches a number of physiologists have examined the
problem in wild-caught and domesticated birds. When fed on a standard
mixture of air-dried seed and kept in temperatures of 22-23°C, domesti-
cated and wild-caught birds consume 24-28% of their body weight in
water over a 24-hour period (Calder 1964; Cade et al. 1965; Skadhauge
and Bradshaw 1974). This is equivalent to about 3.0 mI of water for an
5 Drinking, water relations and temperature regulation 67
average bird (12-13 g). Naturally, water consumption increases dramati-
cally when temperatures increase and Cade et al. (1965) found that when
birds were held at 40°C they drank 6-12 mI of water per day. While
there was significant individual variation in the amount of water needed,
no individual drank 100% of its body weight in one day, but this was
the norm for another estrildine, the Black-rumped Waxbill Estrilda
troglodytes, that was also tested by Cade et al.
Although Zebra Finches find fresh water the most potable they will
drink water from bores with salinities from 0.15-0.3 M NaCl (Badman
1987; Skadhauge and Bradshaw 1974). Oksche et al. (1963) conditioned
wild-caught Zebra Finches to tolera te salinities of 0.7-0.8 M NaCl;
however, domesticated birds could tolerate only 0.6 M which suggests
that the wild birds have a higher maximal renal concentrating ability
(Skadhauge 1981).
Oksche et al. (1963) found that wild-caught birds could be condi-
tioned through gradual reduction in drinking water to survive for at least
several months on 0.5-1.0 mI of water per week at 22-24 oc. Similar
results were found by Priedkalns et al. (1984) and Vleck and Priedkalns
(1985), again using wild-caught birds. These were kept in a room at
27°C with a relative humidity of 40% and fed on standard dry finch
seed, and, while there was no loss of weight, dehydration caused a signif-
icant fall in testis size.
Laboratory birds have be en deprived of water completely yet have
managed to survive on a diet of dry seed only. Domesticated Zebra
Finches held by Cade et al. (1965) lived for more than 250 days without
water, a finding replicated by Lee and Schmidt-Nielsen (1971) who
concluded that their Zebra Finches could survive almost indefinitely
providing there was no heat stress. More than half of Sossinka's
(1972a,b, 1974) birds survived a total deprivation experiment that ran
for 513 days. There was no significant difference in survivorship between
the sexes or between domesticated and wild strains of Zebra Finches.
Body weight initially fell on deprivation, then increased and later
stabilised. Feather condition deteriorated and the bills of a few grew so
long that feeding beca me impossible, and they starved.
Although these laboratory studies in no way suggest that Zebra
Finches in nature can exist without regular access to water, they never-
theless demonstrate an extraordinary ability to endure extremes of thirst
and dehydration that must enable them to sur vive long periods without
drinking. This reserve capacity may be essential to locate new sources of
water when old ones dry up or when seed stocks become depleted. Given
the erra tic nature of the climate over much of the arid and semiarid zone
this capacity should ensure survival.
About a dozen species of granivorous birds have demonstrated the
capacity to survive in the laboratory without drinking water providing
seed was abundantly supplied. These species have a body mass of less
68 THE ZEBRA FINCH
Water economy
Water balance or water flux in birds is an outcome of water intake and
water 10ss and research into this field has been the subject of a number of
detailed reviews (e.g. Dawson 1981; Skadhauge 1981; Webster 1991).
Birds gain water in a 'preformed' state via drinking and vía the moisture
content of the ingested seed. The water cantent of commercial air-dried
seed ranges from 7--10% by mass (Morris 1955; Calder 1965), but
would be much higher for green and half-ripe seeds and less for old, des-
iccated ones. Water is also made by the bird when it metabolises seed.
Metabolic water production (MWP) is highest when carbohydrate is oxi-
dised. Consequently seed, which is rich in carbohydrate and relatively
low in protein and fat, not only supplies the energy needs of the bird but
can produce significant amounts of water. Metabolic rate and MWP
depend on ambient temperature and the range of temperatures where
metabolism in the Zebra Finch is at its minimum (the thermal neutral
zone) is 30-40°C (32-40°C, Calder 1964; 36-42 oC, Cade et al. 1965).
This range is one of the highest recorded in birds and means that when
ambient temperatures fall below about 36°C metabolic rate and MWP
are progressively increased to the extent that below 23°C dehydrated
labaratory Zebra Finches do not need to drink at all beca use MWP
equals the amount lost (MacMillen 1990). H Zebra Finches are not dehy-
drated through gradual water deprivation, ambient temperatures must
drop to 12°C, or lower, befare MWP oHsets that lost from the body.
Thus, the ability of individual Zebra Finches to meet their water needs
from MWP depends on their drinking histories. Tolerance to dehydra--
tion in Zebra Finches may vary clinally from arid habitats to mesic and
humid ones, as well as seasonally within habitats. Serventy (1971) has
observed 'drinking cHnes' in other arid species of Australian birds.
Zebra Finches lose water mainly through evaporation (64%), but also
by excretion and egestion (36%) (MacMillen 1990). Water-deprived
birds are more efficient at reducing losses than non-deprived birds. The
Zebra Finch kidney is highly eHicient at extracting water from urine and
produces one of the most concentrated urines found in birds (2.8 times
more cancentrated than blood plasma) due to much of the nitrogen
being excreted as crystalline uric acid (Skadhauge 1981). A Zebra Finch
dropping is rod-shaped with white uric acid at the anal end of the faeces,
aH of which is surrounded by clear liquid urine (Lee and Schmidt-Nielsen
1971). The faeces are relatively dry due to resorption of water in the
colon and cloaca. The dry weight of heces produced daily by Zebra
5 Drinking, water relations and temperature regulation 69
Finches is 0.3 g but water content (% mass) varies from 80% in individ-
uals not deprived of water (Calder 1964) to 55% in deprived ones (Lee
and Schmidt-Nielsen 1971); these values are among the lowest found in
birds (Skadhauge 1981). Evaporative water loss in the Zebra Finch
occurs mainly through the skin (63%, Bernstein 1971; 48% Lee and
Schmidt-Nielsen 1971); the remainder is lost from the lungs via expired
air. Rates of evaporation relative to oxygen consumption are low in
comparison to other species of birds, and Zebra Finches deprived of
water can make even greater reductions in evaporative water loss by
reducing that lost through the skin but not that lost through pulmonary
evaporation (Lee and Schmidt-Nielsen 1971). A lipid barrier in the skin
of nestling Zebra Finches prevents, almost completely, the passage of
water through the skin; this impermeability is gradually lost after 10
days of age so that by adulthood permeability to water is about 20 to 30
times greater than that of young nestlings (Menon et al. 1988). However,
if adults are depliived of drinking water for up to six weeks the intra-
cellular lipids are re-mobilised and the impermeable barrier is partly
restored so that water loss through the skin is reduced by at least half;
the exceptionally low levels found in the naked nestlings are never
attained (Menon et al. 1989).
Thermoregulation
Mean cloacal temperatures of laboratory Zebra Finches rise with
increasing ambient temperatures and range from 38° to 44°C (Calder
1964; Cade et al. 1965; Bech and Midtgard 1981; Marschall and
Prinzinger 1991). This demonstrates a tolerance of about three degrees
to hyperthermia. If body temperatures reach 45-46°C death occurs
within one hour (Calder 1964; Cade et al. 1965). At high air tempera-
tures, Zebra Finches become completely immobile and initially dissipate
heat passively by sleeking the plumage and holding the wings out from
the body to allow radiation from the thinly feathered sides and axillar
regio n (Figure 5.2). Evaporative water losses also increase aboye 37°C.
When body temperatures reach 42-43°C, all water conservation mea-
sures are abandoned and heavy panting maximises evaporative cooling
(Cade et al. 1965). Schleucher (1993) observed that heavy panting began
in sorne wild populations when ambient temperatures reached 38°C, and
in northern Victoria, birds wiU even begin panting when shade tempera-
tures exceed 32°C. Zebra Finches have an extraordinary ability to dissi-
pate heat through evaporative cooling and can lose up to 1.4 times as
much heat as is produced; this capacity is greater than that known for
any other species of passerine (Calder and King 1963). However, water
losses are so great (0.6 g of water per hour at ambient temperatures of
43°C; Calder 1964), that birds already hardened by long periods of
70 THE ZEBRA FINCH
Fig. 5.2 Panting posture of a heat-stressed Zebra Finch. Note the extended legs, the
sleeked plumage, the open gape and raised carpe/s.
Summary
Zebra Finches are dependent on daily access to drinking water, but sorne
arid populations can live long distances fram watering points, and may
not need to drink every day. Small pools in open areas are preferred sites
for drinking, but Zebra Finches will drink at any location and have
exploited the numeraus bores drilled by Europeans that tap artesian
water acrass much of the eastern arid zone. Zebra Pinches do not drink
by sucking, but use a complex 'double-scoop' mechanism that enables
them to hold the bill tip down in the water when swallowing. The main
advantage is high drinking speed which reduces their vulnerability to
predators; it also enables them to exploit minute quantities of water, and
to exploit sources of water difficult to access. Fresh water is preferred
but Zebra Finches ha ve high tolerances to saline waters in the laboratory
and in the wild. Laboratory birds drink about 28% of their body weight
each day, but this increases steeply with increased air temperatures. In
the laboratory they can survive almost indefinitely without any water
praviding there is no temperature stress and seed is unlimited. Sufficient
water is obtained fram dry seed where small amounts exist in a pre-
formed state and the remainder is produced fram the metabolism of car-
bohydrate found in the seed. Laboratory work suggests that wild Zebra
Finches ha ve an extraordinary reserve capacity to withstand thirst and
dehydration that is only drawn upon during periods of extreme stress.
Zebra Finches have exceptionally efficient means of conserving water;
evaporation through the skin is very low and the kidneys and cloaca
produce extra-dry uric acid and faeces. A high thermal neutral zone
pre-adapts Zebra Finches to high ambient temperatures. Above 43°C
heavy panting maximises evapora ti ve cooling and heat is dissipated very
efficiently, but access to drinking water is essential to resto re the water
72 THE ZEBRA PINCH
The behaviour and ecology of breeding Zebra Finches has been the prin-
cipal focus of field studies to date. Investigations of various duration and
intensity have been made in most parts of the range, except for the
northeastern region.
Colonia lity
Ten Zebra Finch colonies ha ve been studied in detail to date (Table 6.1).
Most studies have concentrated on nesting activity and breeding season-
ality, but only my studies in northern Victoria have monitored seasonal
changes in the size of free-flying flocks.
Nest dispersion
Zebra Finches use nests for breeding and roosting and are one of ten
Australian species of estrildines that build a nest specifically for roosting
purposes (Immelmann 1962a). A resident pair have a regular roosting
nest in a colony which they strongly defend from others in the evening,
but after dark they will eventually allow a desperate individual to enter
and share it. The breeding nest, in contrast, is always defended. Disper-
sion of breeding nests varies according to location, availability and type
of sites, and density of breeding birds. The majority of potential nesting
sites tend to be non-uniform in distribution in most parts of the range.
Consequently, it is unreasonable to make estimates of nesting density for
every location. However, in four colonies, nesting trees were fairIy uni-
form in distribution and densities could be estimated (Table 6.1). The
maximum density found was 76.6 nests per hectare at York, Western
Australia, where a small colony bred in a large thicket (150 m x 20 m) of
hakeas (Hakea preissei), a prickly bush (Immelmann 1962a). The lowest
density estimated was 0.7 nests per hectare for a medium-sized colony
nesting in a low woodland of Long-Ieafed Corkwood (Hakea suberea) at
Alice Springs, central Australia (Zann et al. 1995).
Sorne pairs of Zebra Pinches nest away from the main colony. At
74 THE ZEBRA FINCH
Table 6.1 Size of ten Zebra Finch nesting colonies during breeding and nonbreeding
seasons a
a Breeding season size is based on (i) the number of active nests (eggs or young) during the peak
month of the season and (ii) the number of adults captured that month. During the nonbreeding sea-
son colony size is based on the maximum number of adults observed in winter feeding flocks near
the nesting colony or captured in nets or traps at the colony. In four colonies dispersion of nesting
sites permitted estimates of nesting density (nests per hectare). Dash means no data collected.
b See Figures 2.4 and 3.3.
c Visual estimate of feeding flocks.
d Mistnetted.
'Walk-in trap.
1 Estimated from capture, mark, resighting.
6 Coloniality and breeding ecology 75
York, Immelmann found one pair nesting alone 800 m away from the
others and similarly, at Kununurra, pairs nested on their own from 70 m
to 100 m from the main colony. At the Danaher colony the majority of
breeding birds nested in 12 to 16 small African boxthorn bushes that
grew in a ragged hedge about 190 m long. These formed the main
colony, but other pairs nested in four to seven satellite colonies of one to
two bus hes each that grew at distances of 100 to 1,870 m away from the
main colony (Zann 1994a). A few pairs nested on their own, hiding
them in small bushes several hundred metres from their nearest neigh-
bours. Despite the dispersion of nests at Danaher all the pairs formed
one social group, and there was continuous commuting of members
between all parts of the colony.
Both Immelmann (1962a) and Kikkawa (1980) reported that breeding
pairs at their colonies preferred, where possible, to nest in their own
exclusive bush. Pairs vigorously defended the bush against other con-
specifics, especially those searching for a nesting site. The ground around
the nesting bush was not defended. At Armidale up to 220 separate
bushes and trees were used for nesting by one colony over the four years
it was monitored by Kikkawa (1980), yet the most breeding nests used in
any one month of the breeding season was only 32. By contrast, at my
Danaher colony, multiple nesting in the same small bush was the norm
rather than the exception (Zann 1994a). The majority of nests were one
to two metres apart with up to four or five active nests in the one small
bush being quite common. Pairs defended the nest itself and the nest-
approach perch but persistent newcomers were allowed to build in the
same bush after sorne initial skirmishes. Incubating and brooding birds
regularly interrupted their duties to chase trespassers. Occasionally, two
occupied breeding nests would be in physical contact, but the entrances
faced different directions thus reducing provocation. In the main colony
at Danaher, most nesting occurred in only a third of the bushes while
apparently equivalent bus hes nearby remained empty much of the time.
Favoured nesting bushes tended to change inexplicably from one season
to the next. Multiple nests also occurred in the small bushes that formed
the satellite colonies. Thus, nesting bushes were not a limiting re so urce at
the Danaher colony and this cannot account for the high level of nesting
sociality.
Multiple nests in the one bush have been reported for many locations
in the arid zone where suitable nesting sites are often in short supply at
sorne locations. However, rarely is evidence provided on this aspecto
McGilp (1944) claimed that 21 nests, all of them active, were found in
one Acacia victoriae bush at Oodnadatta, northern South Australia. In
northwestern Australia, Whitlock (1948) found 13 active nests in one
bush and, in central Australia, Immelmann (1962a) reported nine active
nests in a single bush. By contrast, the most nests I found in a single
corkwood tree at Alice Springs in 1986 was three although these trees
76 THE ZEBRA FINCH
Nest predation
It appears that resolution of the forces of attraction and repulsion
between breeding pairs of Zebra Finches clearly differ within and among
colonies across the distribution. Pressure from nest predators appears to
the most important determinant affecting nesting socia lit y in Zebra
Finches. Unfortunately, there are few data on the species of predators
involved, their density and seasonal abundance although most predation
occurs in the breeding sea son. Nest predators at the colonies in northern
Victoria included drago n lizards Amphibolurus spp., Tiger Snakes
Notechis scutatus, Brown Snakes, rats Rattus rattus and mice Mus mus-
culus; Australian Ravens Corvus coronoides and Brown Goshawks were
seen near nests and presumably were responsible for the destruction of
those nests torn aparto At Alice Springs, nest predators included Grey-
crowned Babblers Pomatostomus temporalis, YeHow-throated Miners
Manorina flavigula, Litde Crows Corvus bennetti, T orresian Crows
Corvus orru and Pygmy Mulga Monitors Varanus gilleni. At York,
Immelmann (1962a) found that the Singing Honeyeater Lichenostomus
virescens was the chief nest predator while unspecified species of snakes,
goannas and carnivorous marsupials and rodents were the major preda-
tors at his Kununurra study site in the Kimberley. Barn OwIs Tyto alba
will take roosting adult Zebra Finches; for example, in central Queens-
land 65 Zebra Finch skulls were found in 25 regurgitated pellets at one
roost (P. Woolley, pers. comm.).
At the Danaher colony rates of nest predation were found to be signifi·
cantly lower at bushes in the main colony than at equivalent bushes in
the satellite colonies (Zann 1994a). Re-nesting pairs were significantly
more likely to move to a new bush more than 20 metres away if their
first attempt suffered predation than if it did not; however, they were
conservative in this respect, with many simply re-nesting in the same
bush, sorne even on the same site.
Zebra Finches rarely attempt to repel potential predators; they have
never been observed mobbing in the wild, but something resembling it
has been observed in captivity (Lombardi and Curio 1985a,b). Zebra
Finches attempt to reduce predation by making the nest difficult to find
and/or difficult to reach. If this is not possible their only recourse is to
nest socially in order to exploit the 'selfish herd' effect. This only appears
to be effective when predation levels are moderate. When vegetation of
the nesting bush or tree is dense, nests tend to be placed in inconspicuous
positions and may be dispersed fairly uniformly throughout a colony
thereby making it difficult for predators to find. This strategy appears to
operate at the York, Armidale, Padgett and Cloverlea colonies (Table
6 Coloniality and breeding ecology 77
6.1). Where nesting trees are thinly vegetated, as is the case at Sheppar-
ton and Danaher, and in many parts of the arid zone, nests are almost
impossible to hide and are frequently clumped loosely together, presum-
ably to exploit the dilution effect should they be discovered by a
predator. It is unlikely that colonial nesting would saturate large local
predators.
Irrespective of whether Zebra Finches attempt to hide their nests or
not, all prefer thorny or prickly bushes as a nesting site if these are
available. This hinders but probably does not prevent nests from being
robbed. Where rates of predation are high, Zebra Finches exploit exotic
nesting sites that predators avoid or cannot access. In northern
Victoria, for example, Zebra Finches at several colonies have forsaken
the formidable yet still vulnerable defence provided by boxthorn bushes
to nest in hollow, steel crossbeams that support insulators on power
poles along roadsides (Zann and Rossetto 1991; Zann 1994a). In
addition, most nest predators appear to avoid places of human
habitation, such as farmhouses and outbuildings, and the vicinity
of large raptor nests where they themselves may fall prey. Both sites
appear to attract breeding Zebra Finches and suggests that they
confer lower rates of nest predation (Zann and Runciman 1994). Immel-
mann (1962a) thought that the tendency of Zebra Finches and several
other species of estrildines to nest in sites over water and in the vicinity
of nests of stinging wasps is a further attempt to make nest predation
more difficult.
synchranously than those at York. These birds had one nest only and
simply converted the roosting nest into a breeding nest.
Duration of colonies
In the non-arid parts of Australia, most Zebra Finch colonies are perma-
nently occupied thraughout the year. Numbers, of course fluctuate, but
in the non-breeding season nests are used for raosting and sorne breeding
pairs are found nesting thraughout the long breeding season. Colonies
may persist for a number of years. Kikkawa's (1980) colony at Armidale
existed for at least four years and several of my colonies in northern Vic-
toria have been continuously occupied for at least 10 years.
Size of colonies
The number of pairs breeding in a colony varies thraughout the season
but the number of active nests (eggs and/or young) found in the peak
month of breeding indicates the maximum size of the breeding colony
(Table 6.1). This varied fram seven to 47 nests in the ten colonies studied
to date. The number of breeding pairs does not seem to be determined by
supplies of seed bait used at those colonies with walk-in traps. The num-
ber of adult birds raosting at a colony is difficult to determine during the
breeding sea son since feeding flocks are fragmented and birds may raost
on skimpy platforms or other nest remains that are easily overlooked.
Nevertheless, trapping data show that many more adults occur than
there are breeding nests and suggests that non-breeding birds may be
present at the breeding colonies. These may be dispersing birds in transit,
immigrants about to make a breeding attempt, or visitors from distant
colonies. The seed bait pravided at the trap may be the stimulus that
attracts and keeps sorne of these birds. At Danaher there was no rela-
tionship between the number of pairs nesting each season and the num-
ber of adult and young that arrived fram other colonies (Zann and
Runciman 1994).
After the breeding season, most colonies increase in numbers of adults;
sorne almost tenfold whereas a few colonies fall in numbers. There
appears to be no consistent relationship between trapping method and
numbers of adults in the non-breeding sea son (Table 6.1).
(a)
o Nestlings EL! Young o Adults
1004----------------------------1
90
80
(J) 70
~
.3a. 60
J 50
+-'
¡¡¡ 40
~
~ 30
20
10
O~~~r_~~~--~~~--._~~
Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
(b)
o Adults -+- Breeding attempts
mm +1 SE
90
80
70
.... 60
<V
~ 50
~
~ 40
ni
~ 30
20
10
O~~~~~~~--~~~~~~~
Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul
Fig. 6.1 Annual changes by month in the composition of Zebra Finches at the
Danaher colony over the period 1985-1989: (a) age structure (n = 1,585); (b) mean
number of resident adults (n = 547) and breeding attempts (at least one egg: n =
467 attempts). Residents were those recaptured 228 days after initiaI capture.
Nestlings: 12-20 days of age; young: 20-80 days; and adults >80 days. Birds were
trapped in walk-in traps at least once every month except for winter 1986. Nests
were monitored weekly in the breeding season and monthly in the non-breeding
season.
80 THE ZEBRA FINCH
stage. There were three peaks in numbers of young over the breeding
season: early and late summer and early autumn. Peaks and troughs in
numbers of nestlings and free-flying young coincide, rather than being
offset as one would expect as cohorts move through the age classes. This
is probably due to the fact that the peak s of breeding differ slightly from
one year to the next so that sequences of cohorts are obscured.
The number of adults peaked in mid-autumn, probably due to the
maturation of young of the season. This number feU in late autumn and
early winter due to dispersion and mortality, but peaked again in mid-
spring due to immigration; numbers feU again in the middle of the
breeding season, presumably as fewer immigrants took up residence in
the previous months due to breeding commitments.
Roosting nests
A typical roosting nest is approximately spherical with an outer diameter
ranging from 8-15 cm with walls and roof seldom exceeding a thickness
of 1-2 cm. The side entrance is wide and may have a slight overhang or
awning, but in many cases only half the roof is presento Visitors may be
forced to overnight on an old nest floor or platform. Nests are built of
grass stems or twigs, whatever is locally available. Most roosting nests
are simply the remains of old, dilapidated breeding nests in which the
walls and roof are partIy renovated. Breeding nests in a reasonable state
may be modified for roosting by destruction of the entrance tunnel
and enlargement of the entrance itself (Kikkawa 1980). Purpose-built
roosting nests are smaller and more flimsy than breeding nests with the
number of stems involved in the construction ranging between 180-350
(Immelmann 1962a). Structure of roosting nests varies throughout the
range, for example, neither Immelmann nor Kikkawa found roosting
nests lined with feathers or wool but this was the norm at my colonies
in northern Victoria; furthermore, Kikkawa could distinguish un-
mistakenly a roosting nest from a breeding nest by its wide entrance, but
this was not possible at my colonies in northern Victoria and central
Australia.
pairs and those searching for a nesting site also return early from the
feeding flock. Pre-Iaying pairs may de1ay their departure with the feeding
flock and fly to special 'courting' trees or bushes where they court and
mate in re1ative privacy (Chapter 9). Sorne young and non-breeding
adults may stay in the feeding flock for most of the day, J:>ut after several
hours of feeding there is continual traffic between the colony and the
feeding flock. In the afternoon, social activities predominate to the extent
that nests with eggs or young may be unoccupied for an hour or so, but
eggs with thermocouple implants failed to show noticeable falls in tem-'
perature at this time (Zann and Rossetto 1991). The colony usual1y has
one or two 'social' trees or bushes that serve as centres for resting, preen-
ing, allopreening, singing etc. In southern Australia these sites are she1-
tered from the cold wind and trap the late afternoon sun. The branches
of the social trees have fewer leaves than typical nesting bushes and
allow for comfortable perching of many flock members. Our te1escopes
were often trained on the social tree in order to determine mated pairs
among colour-banded individuals. Immelmann (1962a) also observed
regular bathing and drinking sites with favourite trees for post-bathing
preening, but 1 have not observed these in northern Victoria.
Breeding pairs
Over four breeding seasons at the Danaher colony, breeding attempts of
144 unique pairs (122 males and 123 females and 21 re-pairings) of
colour banded birds were followed. Only 19% of males and 28 % of
r"-
if'MOWW X'f MPNO
I
~"- ~""e" ,~"CW"
'f NRMO X d"PWMN 'f WBMB X d"OBMR
I I
if'UWMP X 'f MBUN 'j'RWMB X if'OWMR
'j' NOBM X d"NUMR
Fig. 6.2 A genealogical tree showing four generations of Zebra Finches at the
Danaher colony, Victoria, over the period 1985-1989. Birds with more than one
partner only re-mated after their first partner had disappeared. Female MBUN, a
resident between 1987 and 1989, appears twice in the tree, having mated with her
grand-daughter's 'brother-in-law'.
6 Coloniality and breeding ecology 83
females were hatched at Danaher, the remainder having immigrated
from elsewhere (Zann 1994a). With such an outbred population few
family pedigrees could be established, despite the potential for rapid pra-
duction of generations. Parents of both partners were known in only six
pairs and in 12 pairs both parents of one partner were known. A sample
pedigree is illustrated in Figure 6.2.
In one pair, parents were brother and sister fram the same clutch.
They hatched in autumn 1988 and made several unsuccessful breeding
attempts with unrelated partners in the spring, but these disappeared,
and the following autumn brather and sister made two breeding
attempts together; the first clutch suffered predation and only one egg of
four in the second attempt hatched. The one nestling fledged successfully.
DNA fingerprinting showed that both brother and sister were full sibs
(Chapter 9). When brather and sister paired it is unlikely that it was a
case of mistaken identity. If cues for kin recognition exist they could
have be en learnt since both siblings were raised in the same clutch; more-
over, the brother had faithfully learnt his father's song and Distance Call
(Zann 1990) so that these diagnostic family traits would have be en avail-
able to the sister at re-pairing.
Pair bonds
Except when incubating and brooding, pairs of Zebra Finches are insep-
arable in both breeding and non-breeding seasons. At Danaher, 65% of
the 144 uniquely colour banded pairs identified making breeding
attempts formed initial bonds, that is, the first bond observed for either
partner; bonds of the other 35% of pairs were either subsequent bonds
for both partners, or a combination of initial or subsequent bonds (Zann
1994a). Breeding pairs that immigrated to Danaher after the age of pair
formation (68% of immigrants) may have had previous partners at pre-
vious colonies.
Fifty re-pairings were detected during the Danaher study. These only
happened when a partner disappeared fram the colony and, presumably,
had died. No bird was ever observed to ha ve more than one mate at any
one time, therefore, social monogamy and divorce by death was the rule.
In 23 cases females re-paired, and in 27 cases males re-paired. These
findings are the first confirmation of the long-held belief that Zebra
Finches are monogamous and pair for life (Immelmann 1962a, 1965a).
Most birds (85%) had one only partner during their term at the Dana-
her colony; a few had two or three and one female had four. There were
no significant differences in the number of mates acquired by the two
sexes (Zann 1994a). Replacement mating could be more common than
this, for in 97 pairs both male and female disappeared simultaneously.
Both may have died together or, more likely, emigrated together, so
nothing is known about how their pair bonds terminated. Since wild
84 THE ZEBRA FINCH
Zebra Finches may live up to five years (Chapter 8) there is ample time
for multiple mates, given the fact that high mortality frequently ruptures
bonds.
Table 6.2 Age of Zebra Finches in days on the date of the first egg of the first
breeding attempt; after Zann (1994a)
-----~----~-_._--
----~.
15
O sameseason
Ol
e (n = 37)
5>- 10
'O
• next season
(n = 32)
Qi
..o 5
E
:::l
Z
O
Sep Oc! Nov Dec Jan Feb Mar Apr
Month of hatching
Fig. 6.3 Number of Zebra Finches by month of hatching that made a breeding
attempt in the sea son in which they were hatched (same) or held over their first
breeding attempt to the following breeding sea son (next). (After Zann 1994a.)
youngster will attempt to breed in the current sea son or defer to the next
season. There will be the c1assic trade-off between costs and benefits of
current reproduction versus survival of the parent and its offspring to the
next breeding season. Foraging skills and seasonal abundance of the wild
seed necessary to feed nestlings should be important considerations in
the decision to defer breeding to the next season or noto At locations
where there was a superabundance of wild seed in late autumn, such as
at Cloverlea, one would predict that the pivotal month for deferring
breeding would be later than that found at Danaher where wild seed was
less abundant. There should be strong selection for parents to breed as
early as possible in the breeding sea son to enhance the possibility of pro··
ducing at least two generations per season, and thus increasing their
reproductive value.
'Dispersal' distance
Sixteen males and 24 females banded in the nest made their first breeding
attempts in the Danaher colony. One male and two females made the
attempt in the same bush in which they were hatched while 15 males bred
in bushes ranging from 165-1,815 m (median 653 m) away; 22 females
bred from 105-1,815 m (median 505 m) from their hatching bush. Differ-
ences in dispersal distance between the sexes were not significant.
Nesting
Nest structure
Zebra Pinches build a typical estrildine breeding nest: a spherical, domed
egg chamber of dead grass stems with a side entrance to which is
6 Coloniality and breeding ecology 87
~----12-24 cm ----_~
inner layer ,
A
3 - 5 cm
outer layer
lining
platform
Fig. 6.4 Structure of a typical Australian Zebra Finch breeding nest.
Nest sites
Given their extensive distribution and range of habitats it is not surpris-
ing that Zebra Finches are the least specialised of all the Australian
estrildines in their choice of nesting sites. Where possible they prefer to
anchor the nest in densely branching thorny shrubs and small trees, sit-
ing the nest in vertical forks or between horizontal branches on the
periphery. In many parts of the range suitable nesting sites are limiting
and almost any site is used, including bizarre ones such as clumps of dry
seaweed on a beach, rabbit holes, holes in cliffs or even the bleached
skulls of catde; nevertheless, they rarely, if ever, nest in grass.
Unless cited specifically the following descriptions of nesting sites are
based on my own observations and those extracted from the RAOU Nest
Record Scheme. In the more setded parts of southeastern and eastern
Australia, Zebra Finches frequently nest in introduced species of trees
and shrubs including fruit trees, (especially Citrus and grape), thorny
weeds (boxthorn, briar rose, gorse, hawthorn, blackberry, brambles and
thistles), and ornamentals (e.g. pinus, cupressus, tamarisk, schinus,
cotoneaster). Of the native shrubs, Melaleucas provide the most common
nesting sites, especially those dense forms, such as M. armillaris, M. lin-
earifolia, and M. styphelioides, which are planted as windbreaks in farm-
land or along roadsides. Thorny species such as Hakea spp. and native
blackthorn Bursaria spinosa are also used for nesting. In swampy areas,
lignum, Muehlenbeckia spp. often provides suitable sites and, in rare
instances, cumbungi Typha spp. will do. In north Queensland, Zebra
Finch nests are frequently found in two exotic thorny weeds, Zizyphus
mauritaniana and Parkinsonia aculeata, and a thorny native, Acacia
farmesiana. In southwest Western Australia dense clumps of the prickly
Hakea pressii were the only sites where Immelmann (1962a) found
Zebra Finches nesting.
The flimsily made nests are not suitable for most species of eucalypts
beca use the sparse branching provides poor anchorage for the founda-
tions and they quickly get blown to the ground. Nevertheless, eucalypts
are used if the nest can be lodged in a hollow or in dense clumps of fruit-
ing bodies or in parasitic mistletoe (Family Loranthaceae). In the Kim-
bedey region Zebra Finches nest in trees of the following species
Eucalyptus papuana, E. camaldulensis, E. dichromorphloa, E.
microtheca, Bauhinia cunninghami, Acacia bidwilli, and Carisa lanceo-
lata (Immelmann 1962a). Immelmann believed that Zebra Finches nest
higher aboye the ground in northern Australia than in other pares
beca use of the greater abundance of nest predators such as snakes and
goannas.
Across much of the arid zone the preferred nesting sites are small
prickly trees of three species, Acacia victoriae, A. tetragonaphylla and
Hakea suberea. In addition, Capparis mitchelli is commonly used in thc
6 Coloniality and breeding ecology 89
Alice Springs region, and Eremocitrus glauca and ApophyIlum anoma-
lum in western New South Wales. Acacia aneura, A. cambagei, Casuar-
ina decaisneana, Atalaya hemiglauca, Santalum lanceolatum, and
Melaleuca glomera are the principal non-spiny species in the arid zone
where nests are found. Zebra finches will also nest in dense, or prickly,
chenopod clumps including Atriplex nummularia, Maireana spp.,
Chenopodium nitrariaceum, Rhagodia spp. and Bassia quinquecuspis.
Where suitable sites are not available Zebra Finches will use those
made by other species including Welcome Swallows Hirundo neoxena,
Yellow-rumped Thornbills Acanthiza chrysorrhoa, Southern Whitefaces
Aphelocephala leucopsis, European Goldfinches Carduelis carduelis, and
Australian Ravens Corvus coronoides. At Kulgera, in the Northern Ter-
ritory, Zebra Finches established a small colony of four nests on the
remains of old mud nests of Fairy Martins Cecropsis ariel placed high up
on the piers of a bridge in a region that provided few suitable nesting
places. Zebra Finches will also exploit nesting hollows excavated in ter-
mite mounds by species, such as the Red-backed Kingfisher Todiramphus
pyrrhopygia. In sorne cases Zebra Finches will not wait until the rightful
owners have vacated their nests and may encounter resistance, with pro-
tracted batdes ensuing (Nielsen 1959).
Zebra Finches are renowned for nesting in the foundations of raptor
nests even where suitable nesting bushes are available nearby and while
the owners are engaged in their own breeding attempts. Nests have be en
found in nest foundations of the following species: Black-shouldered Kite
Elanus axillaris, Black Kite, Whisding Kite Haliastur sphenurus, Black-
breasted Buzzard Hamisrostra melanosternon, Wedge-tailed Eagle
Aquila audax, Litde Eagle Hieraaetus morphnoides, Spotted Harrier Cir-
cus assimilis, and Nankeen Kestrel Palco cenchroides. In southwest
Queensland, 11 Zebra Finch nests were found in a nest of a Wedge-
tailed Eagle, and at least four were active.
It is not uncommon for nesting Zebra Finches to be attracted to busy
human setdements even when natural sites are available nearby. For
example, nests are often found beneath verandahs of houses and sheds,
where they are located in hanging plant containers, in roof spaces and
rafters; nests are also found in hollow fence and gate posts, in pulley
block s on moored fishing trawlers, in insulator cross-bearers on power
poles, up irrigation pipes, and in tractors, old car tyres and car bodies.
Even the fuselage and engine cowlings of newly parked aircraft on
airstrips have provided nesting sites. At all these sites the birds fearlessly
go about their business, especially during the building stage, and tend to
ignore most human presence. North (1909), for example, describes a
pair nesting successfully amid the din of a busy blacksmith's shop. Of
course, they become furtive and wary should anyone attempt to make
detailed observations.
90 THE ZEBRA FINCH
Clutch size
Zebra Finches laya modal clutch of five eggs, but range in number from
two to seven (Figure 6.5). Only three studies of clutch size have been
made to date. At Griffith, New South Wales, Frith and Tilt (1959) found
a mean of 4.7 :±: 1.1 (s.d.) eggs in 221 clutches examined over three
breeding seasons. Clutch sizes were significant1y smaller at the start and
120
100
80
>-
ü
e 60
Q)
::l
U
Q)
.... 40
LL
20
O
2 3 4 5 6 7
Clutch size
Fig. 6.5 Number of eggs la id per clutch in Australian Zebra Finches at the Danaher
colony in northcrn Victoria over four brccding seasons.
6 Coloniality and breeding ecology 91
end of the breeding season, presumably when conditions were not opti-
mal for breeding. At the Padgett colony in northern Victoria, I found a
mean of 4.85 ± 0.83 eggs (n = 112), which was not significantly different
from that at Griffith. These Padgett clutches carne from the second haH
(January to April) of the 1976-1977 breeding season; unfortunately, no
nest inspections could be made in the first half of the season. However,
at the nearby Danaher colony the mean clutch size over four breeding
seasons (1985-1986 to 1988-1989) was 5.04 ± 0.98 (n = 303), and was
significantly larger than those found at Griffith and Padgett (Zann
1994a). Clutch sizes in the autumn months (March and April) were sig-
nificantly smaller than those for the preceding six months of the season.
A seasonal decline in avian clutch size, such as this, is typical of bird
populations and three non-exclusive hypotheses have been proposed (see
review by Rowe et al. 1994):
(1) sub-optimal parental condition;
(2) a decline in offspring value due to a lower probability of survival of
late-hatched young, which in turn, leads to
(3) an increase in the cost of reproduction.
While it is unlikely that hypothesis (1) would apply to experienced
breeders it is possible that it could apply to no vice breeders, who them-
selves hatched earlier in the season (see below).
Birds at Danaher were trapped at a walk-in trap baited with seeds ad
libitum, so it is conceivable that the extra seed was responsible for the
enlarged clutch size-it is known in other species that experimental sup-
plementation of food can improve female condition and lead to an
increase in clutch size (H6gstedt 1981; Newton and Marquiss 1981;
Dijkstra et al. 1982). Nevertheless, no significant differences were found
in the size of clutches laid in the second half of the sea son among the
Danaher, Padgett and Griffith birds (Zann 1994a). Thus, ad libitum food
may not necessarily infla te the clutch size, at least in the second half of
the season in northern Victoria. Brood size (and presumably, clutch size)
was not significantly affected by experimentally induced differences in
net energy gain among breeding domesticated pairs of Zebra Finches
(Lemon 1993; Lemon and Barth 1992).
In experimental manipulations of diet of domesticated Zebra Finch
nestlings, Haywood and Perrins (1992) discovered that clutch size was
permanently affected by the level of nutrition experienced by a female dur-
ing her first month of life. The amount of food, rather than its type or
quality, caused the effect (see Chapter 7). Thus, female nestlings that had
high rates of food consumption grew faster than those with low rates, and
subsequently laid larger clutches of eggs even though both groups of
breeding adult females were kept under identical conditions of ad libitum
food. Although samples sizes in this experiment were small (n = 12), this is
92 THE ZEBRA FINCH
Laying times
In common with most species of birds Zebra Finches lay one egg each
day, mostly around sunrise. To determine this, we inspected new nests at
the Danaher colony every two hours, from before sunrise until 15:00 hr
Eastern Standard Time during January and February 1991. Of 31 eggs
from 16 elutches, 27 (87%) were laid in the two-hour period starting
just before sunrise. No eggs were laid after 13:00 hr. Domesticated
Zebra Finches also lay their eggs in the earIy morning, and, coinciden-
tally, Haywood (1993a) also found that 87% of eggs were laid within
the first two hours after the cage lights were turned on.
The first egg is frequently laid before the nest is complete. Frith and
Tilt (1959) found that the average time between the onset of building
and the first egg was six days with building being completed in another
seven days. In both central Australia and northern Victoria, the first egg
may be laid with only the floor and lower walls of the nest chamber
standing and this may only take two days to constructo In most
instances, the nest is fully constructed before the first egg is laid, but lin-
ing is not added until incubation begins.
Eggs
Eggs are matt white in colour when newly la id but become polished with
age. At the narrow end faint greyish streaks or mottling may occur in eggs
laid by certain females. Only preliminary studies of egg size have been
made on the Australian Zebra Finch. Dimensions and shape of eggs laid
by the Australian subspecies vary considerably, even when food is not
limited. In my aviaries, eggs freshly laid by second and third generation
offspring of wild-caught Zebra Finches weigh about one gram (1.0 ± 0.10,
range 0.75-1.25 g, n = 147) and laying sequence had no significant effect
on weight (F1,145 = 0.44, P = 0.51), nor was there any difference between
fertile and unfertile eggs (t145 = 0.27, P = 0.79). However, there was a
significant female effect (Welch test, F26,31 = 9.49, P<O.OOl) indicating
94 THE ZEBRA FINCH
Re-nesting
The mean (± s.d.) number of clutches per season for males at the Dana-
her colony was 1.9 ± 1.16 (range 1-6, n = 127) and 1.7 ± 0.91 (1-4, n ::::
127) for females. Within a breeding sea son the mean interval in days
6 Coloniality and breeding ecology 95
between the dates of the first eggs of successive clutches was significantly
shorter when the previous attempt failed (33 ± 11 days, n = 35) than
when it was successful (52 ± 16 days, n = 48) (Zann 1994a). The short-
est time between re-nestings was estimated at 34 days, when a female
was found incubating a full clutch while her mate cared for two
fledglings about 22 days old. Within a single breeding season at Dana-
her, 54% of females (n = 123) laid a single clutch only, 25% laid two
clutches, 14% three, 4% four, 1 % five and 1 % six. These are minimum
values beca use sorne attempts may have gone undiscovered, and attempts
made at other colonies before, or after the breeding pair arrived at Dana-
her, were likely.
45
Day 12
I/l
;:¡
40
(f)
Q)
O
(f)
<D
~
Ol
<D 35
O
30
W3 W5 W6 W7 W8 W9
Pair
Fig. 6.6 Day and night egg temperatures during incubation for six pairs of free-living
Zebra Finches at the Danaher colony in northern Victoria. Bars show means and
one standard error. The number of hours with 12 five-minute readings is given
above the bars. (After Zann and Rossetto 1991.)
6 Coloniality and breeding ecology 97
tion temperatures of wild-caught Zebra Finches (n = 7 pairs); daytime
ones were significantly lower than those at night but were higher than
daytime temperatures of free-living birds. Nest attentiveness accounts for
these differences-it is greater at night than during the day and less
among free-living Zebra Finches than wild-caught ones. We found no
significant differences in egg temperatures between early and late stages
of incubation, unlike what has been found in other small passerine
species (Webb 1987). Temperatures were also maintained or even
increased after hatching.
Vleck (1981) found that domesticated Zebra Finches had a mean egg
temperature during incubation of 37°C whereas El-Wailly (1966)
obtained a lower temperature of 35.rc. The incubation temperature of
most species of birds is around 35°C (Kendeigh et al. 1977).
10
3 6 9 12 15 18
Fig.6.7 Mean daily egg temperatures befare, during and after incubation in a wild
caught pair of Zebra Finches. Each data point is a mean of 24 x 12 readings logged
every five minutes. Full incubation began after the fourth egg was laid (day 4) and
full incubation temperatures were not reached until the fifth day. The Brst egg
hatched on day 14. (Aftcr Zann and Rossetto 1991.)
cooling: most eggs still hatched despite being cooled down to 18°C for
periods of 10 hours and were much more resistant to cooling than other
species tested (estrildines, fringillids and ploceines).
Hatching
Hatching of the clutch normally spans two days in wild Zebra Finches, but
is quite variable. Weekly nest inspections made late in the day at Danaher
over four breeding seasons yielded 144 nests in which the first young were
believed to have hatched on inspection day. In sorne clutches aH six eggs
were estimated to have hatched on inspection day, that is total synchrony,
but the majority (47%) of clutches had only 50% of eggs hatching on
inspection day (Figure 6.8). During a subsequent study of egg dumping,
we made daily inspections of 30 nests at Danaher and found that aH
clutches hatched within two days and that 60% had more than 50% of
the clutch hatch the first day. This proportion is not significantly different
from that estimated from weekly inspections (G l = 1.63, n.s.). These find-
ings are consistent with those of Immelmann (1962a) who found that, in
his colonies in Western Australia, two young hatched in the morning,
another one or two around midday, and depending on the size of the
clutch, another one or two the foHowing day. Thus, hatching in wild
Zebra Finches normaHy spans more than one day, which, by definition, is
considered asynchronous hatching (Amundson and Slagsvold 1991).
Hatching in most pairs of wild-caught Zebra Finches is less syn-
chronous than that of free-living birds. Eggs tend to hatch one each day,
40
•
O
weekly inspections (n = 144)
-
~
o
'-'
(/)
Q) 20
~
....::J
()
Ü
10
o
<20 21-40 41-60 61-80 81-100
Development of young
Mouth-markings
Like all estrildines, Zebra Finch nestlings have a conspicuous species-
specific pattern of black markings in the mouth, which are prominently
displayed during begging when the gape is opened wide (Figure 6.9).
Whereas the mouth markings of the Lesser Sundas Zebra Finches have
not been described, those of the Australian subspecies have been studied
in detail. There are five sets of marks in the Zebra Finch:
(1) twin marks that form a chevron in the tip of the premaxilla;
(2) five roundish spots that form a 'domino' pattern in the centre of the
upper palate;
6 Coloniality and breeding ecology 101
(3) two lateral tongue spots towards the tip of the tongue;
(4) a 'half-moon'-shaped mark in the sublingual regio n of the lower
mandible; and
(5) twin tear-shaped marks each side of the gape flanges.
The background colour of the upper and lower palate is off-white with
yellowish tints in sorne individuals. The tip of the tongue is white, the
remainder, a pink flesh colour. A swollen white-pigmented flange forms
outside around the beak angle and contrasts with the twin tear-shaped
marks when the gap e is open wide. There is considerable variation in the
darkness and shape of the marks, but les s within than between broods.
All marks begin to fade around fledging time. This is fastest with the pre-
maxilla marks and slowest with the domino spots which may persist
beyond 50 days of age and even after day 100 faint smudges can be seen.
Domesticated strains of Zebra Finches often fail to develop mouth
markings. While the wild-type has markings similar to those of wild
birds, the white-morph has no black marks in the mouth at all and the
beak flange lacks the white pigmentation. The completeness of markings
in other colour morphs varies between these two extremes (Muller and
Smith 1978). Experiments showed that parents of both wild-type and
white morphs prefer to feed wild-type nestlings, which led Immelmann
et al. (1977) to conclude that the mouth markings were primarily
responsible for stimulating the feeding response in the parents. More-
102 THE ZEBRA FINCH
over, they predicted that discrimination would be stronger among free-
living birds where food would be limited so that young with mouth pat-
terns that deviated even slight1y from the species-specific pattern would
receive less food and suffer higher rates of mortality. Indeed,
Skagen (1988) found that nestlings of domesticated Zebra Finches with-
out palate markings had significandy slower growth rates than those
with markings under conditions of food shortage, but there was no
difference when food was abundant. It would be interesting to test
whether Australian and Lesser Sundas parents can discriminate young in
mixed broods.
The mouth pattern in the Double-barred Finch is similar to that of the
Zebra Finch, except that the upper three spots of the domino are elon--
gated into thick lines and the pigmentation between the two lower
domino spots is more intense. In the Long-tailed and Black-throated
Finches the three upper domino spots are elongated still further so that
they are fused together to form one mark of a 'horseshoe' shape. Steiner
(1960) and Wolters (1957) believe that the five-spot domino pattern is
the most primitive form, occurring in most African estrildines and in
more than half of the Australian ones, whereas the horseshoe pattern is
the derived formo Immelmann (1965a) pointed out that this evolutionary
transition appears to have occurred independently in both African and
Australian species among distantly related groups. Inter-specific brood
parasitism is postulated as one factor that has led to the origin and main-
tenance of mouth patterns in the Estrildinae (Nicolai 1964; Kunkel
1969). Furthermore, it is feasible that intraspecific brood parasitism,
which can occur at fairly high levels in Zebra Finches (below) may be
responsible for maintaining variation in mouth patterns between broods
and possible discrimination by parents. Experiments are needed to test
this idea.
Morphological changes
My observations on nestling development of free-living Zebra Finches
agree with Immelmann's (1962a) description. On hatching day (day O)
nestlíngs are naked and pink-skinned with a few pale-grey down feathers
on the head and back, but by day 2 slate grey pigmentatíon appears in
the middle of the back and quickly spreads laterally covering the flanks,
thighs and top of the head. Pigmentation of the dorsal surface of the
wings begins by day 3, and the whole dorsal surface, flanks and outside
thighs are pigmented by day 5 or 6. The bill of newly hatched nestlings is
a light-horn colour, but on day 2 and 3 dark pigment begins to spread
on upper and lower mandibles and by 10-12 days of age pigmentation is
complete and both mandibles are uniformly black. In the white morph
young do not develop black pigments on the bill, but by fledging time the
horn-colour gradually changes to a pale pink (Immelmann 1959).
6 Colonia lity and breeding ecology 103
The eyes begin to open around 6-7 days of age, forming a narrow slit
at first and become fully open by day 10-11. Thin hairs precede the first
pins of the remiges which break through the skin on day 6 and grow to
about 3 mm by day 10, thereafter increasing by about 3 mm per day
until day 18 when they are about 28 mm. The sheath of the pin splits
open around day 11 exposing the dark vanes of the feathers. The first
pin of the rectrices appears on day 9 reaching 5 mm in length by day 12
and 13 mm by day 18. Growth of the rectrices and remiges is faster than
that of the Long-tailed Finch, but slower than that of the Crimson Finch
(Immelmann 1965a). Pins of the contour feathers appear on the dorsal
tracts around day 7 and those of the ventral tracts appear the next day
leading to complete coverage of the body by around day 14. Down
feathers are gradually lost, first from the shoulders, then the back, the
thighs and lastly the back of the head.
In the wild, nestlings are remarkably resistant to cold and exposure
(Immelmann 1962a). Nestlings as young as four or five days will survive
without food and endure near-freezing temperatures for periods of up to
36 hours. Immelmann believes that they enter sorne form of torpor
where digestive processes cease and food remains in the crop; warmth
resto res them.
On hatching day, the crop of the newly hatched nestling is empty as
parents normally do not feed them until the next day, but sorne give
small amounts of fine, green seeds which sit on the right side of the crop,
which fills first. After day 3 the left side of the crop also stores seed, but
in smaller amounts than the right side; equal amounts are found in both
sides by day 5 if large quantities need to be stored.
Growth rates
Unfortunately, growth rates of nestlings have not been studied in free-
living Zebra Finches, but have been investigated in domesticated birds
in a number of studies. Sossinka (1970) found no significant differ-
ences in the growth of aviary-reared nestlings of domesticated and wild-
caught birds. Like most estrildines investigated to date, Zebra Finch
nestlings are slow developers by comparison with other granivorous
species: they spend longer in the nest and take longer to reach adult
weight than fringillids and ploceines (Ziswiller 1959; Lack 1968;
Sossinka 1972b).
Nestlings of domesticated Zebra Finches weigh between 0.6-0.9 g at
hatching and gain on average 0.66 g per day for the first 10 days; they
weigh on average 4.6 g at day 7 and 8.0 g at day 12 (Skagen 1988) and
9-10 g at fledging in both sexes (Burley 1986c). Newly fledged Zebra
Finches at the Danaher colony weigh around 10 g and reach their adult
weight at a median age of 51 days (range 26-129 days, n = 100), with
those attaining a heavier adult weight taking longer. In domesticated
104 THE ZEBRA FINCH
females, Sossinka (1980a) found that length of body, wing and bill reach
adult size between 27 and 29 days of age.
The energy demand of a Zebra Finch nestling increases steeply each
day from hatching to day 8, where it peaks at 8.8 kJday-\ after which
demand decreases as increasing feather cover assists thermoregulation
(Lemon 1993).
The quality and quantity of the nestling diet affects growth rates and
the size at adulthood. Boag (1987) found that a high protein diet
increased the growth rate of most morphological characters he mea·-
sured, affecting weight and tarsus length the most, while bill and
plumage were affected least. Weight at adulthood was also affected by
the quality of the diet with differences reaching 7-10%. Skagen (1988)
found that the growth rate of nestlings was significantly faster when food
was abundant than when it was limiting, and that variation in rate of
growth was greatest between nestlings where food was limited. Nestling
weight at age 12 and adult weight was positively correlated for both
males and females. This effect of food limitatíon on nestling growth rates
and final adult weight was confirmed by Haywood and Perrins (1992)
who also found permanent detrimental effects on egg and clutch size.
When food is not limiting, but the net energy intake each day is
reduced by high foraging costs, weight of nestlings at fledging is signifi-
cantly reduced and the growth rate slowed significantly (Lemon 1993).
The task remains to see if similar relationships can be established in wild
populations and in particular, the relationship between egg size, nestling
size, and size at first breeding.
Sexual maturation
Sossinka (1970, 1972b, 1975, 1980a) investigated sexual maturation of
both sexes of Australian Zebra Finches. There were no significant differ-
ences between the sexes in body growth, changes in colour of the bill,
and onset of post-juvenal moult. Consequently, these changes do not
appear to be controlled by sex-specific factors. However, development of
the gonads proceeds at a different rate-growth is logarithmic in males
and linear in females.
Testis volume increased dramatically, reaching adult levels around 70
days--its rate of growth was not affected by day length (LD 8:16 vs.
16:8). Spermatogenesis proceeded with corresponding swiftness, the ear-
liest mature sperm appearing at 59 days, and by 70 days most individu--
als had fully developed sperm. Testosterone production matched testis
growth. There was a brief juvenile refractory period between day 18--34
where testis size, spermatogenesis and testosterone production leveled
off, forming a plateau on the steep growth curves (Prove 1983). This rate
of sexual development is rapid, even for small tropical species. For
6 Coloniality and breeding ecology 105
example, testis development in the Spice Finch Lonchura punctulata
took around seven to eight months and the juvenile refractory period
lasted over three months (Sossinka 1975). Sexual maturation is even
more rapid in offspring of wild-caught Zebra Finches raised under
identical conditions. By 44 days of age the testes were almost twice the
volume of those of domesticated birds, and the stage of spermatogenesis
significantly more advanced (Sossinka 1970). Sexually dichromatic
plumage marks were also less advanced in domesticated Zebra Finches,
but there were no significant differences in body size.
Sexual maturation in domesticated females is slower than that of
males. Sossinka (1980a) measured the weight of the ovary and the
largest follicle from day 10 to day 100. The diameter of the largest
follicle increased linearly until 100 days of age when it reached 1.0-1.2
mm, after which it ceased growing. It only started to grow again if
breeding conditions arose, increasing exponentially for a period of about
two weeks, during which ovulation occurred and, ultimately, the first egg
laido Sossinka termed the first growth period the 'stage of primary
development', and the second, the 'stage of stimulated development'. If
breeding conditions arose during primary development there was no rest-
ing stage and eggs could be laid as early as 90 days of age. However,
maturation must be faster in free-living birds since females as young as
62 days laid their first egg in the Danaher study (Zann 1994a). Sossinka
(1980a) also found that only a few follides reached the resting stage in
the first 100 days of life after which other follides began to undergo pri-
mary development. Possibly, young females rush through the develop-
ment of a few follides in order to exploit any breeding opportunities that
might arise before the ovary is fully developed, since the weight of the
ovary did not reach that of adults until at least 250 days of age. No dif-
ferences in clutch size were found between young females and adults in
the Danaher study, so it is conceivable that more follicles undergo pri-
mary development in the first wave in free-living females than in domes-
ticated ones. Another interesting finding made by Sossinka was that
females held under short day length (LD 8:16) too k significantly longer
to grow ovarian follicles during primary development than those held
under longer day lengths (LD 16:8).
Breeding success
Pooled data over four years showed that 35% of breeding attempts,
41 % of incubated eggs and 54% of incubated clutches produced at least
one 'fledgling' (Zann 1994a). These estimates are slightly inflated
beca use the number of fledglings produced was based on the number of
nestlings of 12-14 days of age, not fledglings observed outside the nest.
It is possible that sorne may have fallen prey before they could fledge.
However, young are quite mobile after this age and may escape nest
predators. For example, when forced to fledge, young explode out of the
nest amid a series of loud identity calls and sorne can fly 10-20 metres
before hitting the ground, whereupon they freeze and are difficult to
locate. y ounger birds flutter to the ground beneath the nest and move
rapidly to any dark nook or crevice at the base of the bush and freeze.
Breeding success at Danaher was similar to that found by Kikkawa
(1980) at Armidale, however, but fewer clutches produced young than in
the population studied by Frith and Tilt (1959) at Griffith (Table 6.3).
Success of breeding attempts at the Danaher colony was low by comparison
Table 6.3 Breeding succcss of Zebra Finches at three locations in eastern Australia
based on four seasons' data
" Per cent of completed clutches that f1edged at least one young. Frith and Tilt (1959) are not
explicit (p. 291): 'Of 221 clutches begun in the four years of study, 172 werc completed, and of
these 128 ultimately produced young.' 1 assume these young left the nest.
b Percentage of eggs laid that resulted in a f1edgling.
6 Coloniality and breeding ecology 109
to that of eight species of Australian passerines recently summarised by
Ford (1989).
The number of successful breeding attempts and incubated clutches at
Danaher differed significantly across the four breeding seasons due to the
larger number of failures due to predation in the 1988-1989 season.
Overall, breeding success was significantly greater in the second half of
the breeding season than in the first half (Zann 1994a).
Of 626 young banded just before fledging, 197 (31.5%) were recap-
tured at the walk-in trap as free-flying individuals ~ 35 days of age. This
is after the age of nutritional independence and is presumably the num-
ber of fledglings that have survived to independence, since limited flying
ability prevents dispersal before this age. At 80 days of age, the time of
first breeding, 137 young were recaptured (69.5% survival), but sorne
would have dispersed and would not have been available for recapture.
Thus, at least 9.0% (137/1526) of incubated eggs produced young to
breeding age (Figure 6.10).
370 eggs
depredaled
(24%)
65 died (9%)
( slarvalion/accidents)
Fig. 6.10 Schema summarising breeding success of Zebra Finches at the Danaher
colony in northern Victoria based on four seasons data (1985-1989). A breeding
attempt was defined as one where there was at least one fresh egg. The number of
fledglings was based on the number of nestlings banded in the nest around 12-14
days after hatching.
6 Coloniality and breeding ecology 111
Table 6.4 Breeding success of Zebra Finches by size of clutch in nests that did not
suffer predation (from Zann 1994a)
Clutch Number of Number Number Per cent Per cent Per cent X fledged per
Slze clutches of eggs of eggs eggs hatchlings eggs breeding
hatched hatched bandeda banded b attemptC
2 2 4 3 75 100 75 1.5
3 9 27 26 96 100 95 3.0
4 52 208 185 89 92 80 3.3
5 93 465 369 79 91 66 3.8
6 66 396 326 82 87 67 4.6
7 8 56 49 88 97 81 5.8
n 230 1156 958
Nest failures
Sixty-six per cent of aH attempts (310/471) failed because of nest preda-
tion. Early in the nesting cycle, 36% (168/471) of attempts failed before
the clutch size could be determined and another 47% (142/303) of incu-
bated clutches were subsequently killed before young could be banded
(Figure 6.10). In sorne seasons (e.g. 1985-1986) there were higher rates
112 THE ZEBRA FINCH
of predation in the first half of the season than in the second, whereas
the reverse was true in other seasons (e.g. 1987-1988). Overall 17%
(198/1156) oi incubated eggs not taken by predators failed to hatch
because of inadequate parenting. Failures were greater in the first half of
the eight month breeding season than in the second half. Nine per cent
(65/691) of nestlings failed to reach banding age due to starvation or
accidents and the percentage of these failures varied significantly over the
course of the study and increased pragressively from the first season to
the lasto
Summary
Australian Zebra Finches nest in colonies throughout the year; these
range in size fram just a few pairs up to 40-50 pairs. Nests are located in
thick, usually thorny, bushes close to feeding sites. There are pranounced
annual changes in abundance and age of individuals present in colonies.
Nesting density varies according to habitat and geographic location and
ranges fram 0.7 to 76 ha-l. Territorality is confined to the nest itself, but
an entire bush may be defended initially. In sorne colonies breeding nests
may rest in contacto Nesting socia lit y reduces predation when predator
levels are low to medium, but provides no protection when rates are
high. Corvids, raptors, snakes, goannas, rats and mice are the main nest
rabbers. Zebra Finch colonies ha ve a degree of functional structure with
regular social sites, bathing sites, and arrival and departure sites.
Zebra Pinches raost in nests and under sorne conditions this reduces
loss of body heat. Breeding nests are slightly larger than raosting nests,
and have a long side-entrance tunnel; they are made of grass stems or
twigs, and are highly variable in dimensions, and construction materials.
Breeding nests are lined with feathers and pravide an impraved micra-
climate during cold weather in southern parts of the range, and prabably
insulate eggs and young from excessive heat in summer. Eggs are laid
one each day, just after dawn. lntra-specific braod parasitism (egg dump-·
ing) ranges from 13 % to 32% of braods and may be a means of reduc--
ing the effects of nest predation. The modal clutch size is five eggs, with
smaller clutches in the last months oi the breeding season. Clutch size
does not vary geographically, nor among successive clutches within
females, nor between clutches laid by experienced and novice females.
Both sexes incubate, but only the iemale develops a brood patch and
incubates at night, nonetheless, the male is equally competent as the
female in heating the cggs. The average incubatíon temperature ís 36°C
and begins after the fourth egg is laido Incubation bouts range fram 35 to
190 minutes and may depend on commuting times to and from foraging
sites. Eggs take 11-15 days to hatch, depending on levels of incubation
attentiveness by the pair; the minimum incubation time is 11.25 days.
6 Coloniality and breeding ecology 113
Hatching of the clutch spans around two days in the wild, but in captiv-
ity one egg may hatch each day so that young are of different sizes.
y oung Zebra Finches have species-specific palate markings that aid the
release of the parental feeding response. Quality and quantity of rearing
food affects growth rates of nestlings, final adult size, and future egg and
clutch size. Laboratory males atta in testes of adult size and can produce
sperm around 70 days post-hatch and females can ovulate before 100
days of age, but wild birds mature even earlier.
Zebra Finches are socially monogamous, and pair for life, but re-
pairing is frequent due to high mortalities. In northern Victoria, young
hatched in the first half of the breeding season make up 44% of pairs
making breeding attempts in the second half of the season. For birds
breeding in the season of hatching the median age of first breeding is 95
days in males and 92 days in females. There were no significant differ-
ences in breeding success between novice and experienced pairs. High
nest predation limits breeding success: 35% of nesting attempts, 41 % of
incubated eggs and 54% of incubated clutches produce at least one
fledgling. Only 9% of eggs produce young to breeding age.
7 Breeding periodicity
'Why this hardy little finch should be so much more abundant than its
congeners is a puzzle; but no other is so prolific, for 1 do not think that at any
period of the year nests might not be found if searched for.'
F. Lawson Whitlock, Western Australia (in Cayley 1932).
"t::1 Alice
,'lh 'T"" ,
1
0 •• , • .". -
JFMA~JJASOND
/if.L
:1i - \ 30
r:ill.1 .1 ••
1
I n=21
'T!:'~llh
",,,; i~\
! r JFMAM JASOr--O
(j)
.!
j
Goldffelds • cC1S/?
,~, ~ ;~~,~".) ~c\0~;} ']IIII:~'" ...,,1
G Armidale ( ¡)
~ 20~
-r. 1I .1
n=88 1 FMAMJ.)ASONO
~
,1
z o"
1
10
~ ! I~! "
; ASQND o
- - - - - - - • 'Western New South Wales'
::.j'~"
30
20
: o~
-r """'1""1" .. "'1""
l' (h)
,
i
I
MAMJ.JASOND
(1) 0Wunghnt.:
3°1 í
r:~I. ••
JFMAMJ~:ASOND
n=468 , 1II1 '"
o
Z
7 Breeding periodicity 117
Fig. 7.1 Timing of breeding of Australian Zebra Finehes. Graphs show annual breeding
at five loeations (a, d, f, g, and i) based on monthly monitoring of speeifie populations.
Other graphs (b, e, e, h, j, k, and 1) are eompilations of regional breeding data taken from
the RAOU Field Atlas of Australian Birds (1977-1981) and the RAOU Nest Record
Se heme (1957-1992). The y-axis shows the per eent of nests with eggs diseovered eaeh
month exeept for graph (a) whieh shows the mean number of young hatehed eaeh month
based on baekdating of hatehing date from aging of free-f1ying young eaught at a walk-in
trap. (a) Aliee Springs-central ranges: 1986-1992 (From Zann et al. 1995); (b) 'Pilbara':
0
loeations between 20 -25°S and 116°-126°E; (e) 'Goldfields': loeations between 25°-33°S
and 115°-126°E (excluding York and the wheatbelt distriet of Western Australia); (d)
York (31°52'S, 116°47'E): 1959-1964 (Davies 1979); (e) 'South Australia': loeations
0
between 2r-35°S and 133°-140 E; (f) Wunghnu: 1985-1989 (Zann 1994a); (g) Griffith:
1954 (Frith and Tilt 1959); (h) 'Western New South Wales': loeations between 30 0 -35°S,
139°-145°E; (i) Armidale: 1961-1965 (Kikkawa 1980); (j) 'East Coast': loeations between
21°-33°S, 145°-153°E; (k) 'North Queensland': loeations between 17°_21°S, 139°-147°E;
(1) 'Western Queensland': loeations between 20 0 -26°S; 139°-145°E.
(a)
1oor---------------------------------------------~
"O
<D
i3 75
(¡j
oC
'- 50
<D
.o
E
:l 25
Z
(b)
250~------------------------~------------------~40
-t-'+,'t- .*,.*-:.,,... ':+-"'i"''*''i.
2oo \ ,j'-* ',,+,"1- ;1" , a;:
30 ?<
:¡"'''!-.l \./ SI<>
\\., ,1 ó P'q, \.,,....¡.,.f
~ 20 3
:;'
,0-0' b~
Ó ~ ro
ó' 'o. 10 3
"O
o
O~.,~~~~YT~~~~~~~~~~~~~~~~~ O o
JMMJ SNJMMJ SNJMMJ SNJMMJ
25
o 1985/86 (n = 1 OO)
-en 20
o.
E 15
<1>
• 1986/87 (n=64)
~ 10
5
O~.-~~~~u-~~u-
Aug Sep Oct Nov Dec Jan Feb Mar Apr May
(b)
50~------------------------------1
40
230
o.
E
~ 20
;{
10
0~~~ll4~~~~~~~~~~~
Aug Sep Oct Nov Dec Jan Feb Mar Apr May
Fig. 7.3 Breeding attempts (~ one egg laid) of Zebra Finches at the Danaher colony
over four breeding seasons by month (a) 1985-1986 and 1986-1987, (b) 1987-
1988 and 1988-1989. (From Zann 1994a.)
Although summers are not particularly dry, the region is irrigated from
about October to April to compensate for the irregularity in rainfall and
the high evaporation. Data shown in Figure 7.1f are from one small
(Danaher) colony at Wunghnu and show breeding in all but the four
winter months. Each of the four seasons data show a spring peak in
breeding, but the timing and extent varied from year to year (Figure 7.3).
In three of the four breeding seasons there was a significant positive cor-
relation between the number of breeding attempts made each month of
the season and the mean minimum temperatures for the current month
(Zann 1994a).
Monthly breeding activity at Padgett and Cloverlea, 3.7 and 6.7 km
west of Wunghnu, respectively, showed more pronounced autumn peaks
in breeding. These were as large as those in spring, or larger (see Figure
4.2). A detailed study of diet and breeding conducted over 15 months at
Cloverlea in 1981-1982 showed that each surge in nesting activity over
the breeding season corresponded with a new flush of ripening grass
seeds. The more abundant and larger the species of seed the greater the
breeding response. The timing of the seed flushes depended on the pre-
vailing temperature and the seeding phenology of each species.
The pattern of breeding in northern Victoria corresponds well with
122 THE ZEBRA FINCH
Nix's (1976) plant growth response model calculated for Deniliquin, a
region 80 km north-west of Wunghnu. This model has a bimodal distri-
bution with a peak in spring and a smaller one in autumn. The indige-
nous grasses have comparable seeding phenologies to those of the exotic
pasture grasses that have largely replaced them at this Iocatíon so that
breeding from early spring to late autumn would have also been possible
before European settlement.
South Australia (Figure 7.1e)
Southeastern parts of the State of South Australia have a humid Mediter-
ranean-type climate with wet, cold winters and dry summers. Spring is
the time for maximum plant growth. As one progresses northward,
precipitation deereases rapidly and its oceurrence becomes increasingly
episodic and irregular such that arid conditions prevail; consequently,
there is no regular annual growth periodo The 188 breeding records from
the Nest Record Seheme come principally from the more populated
eastern half of the state below latitude 30 0 S. They show a pronouneed
peak in spring that oeeurs in August and September in the arid northern
parts and in October in the semiarid southern region.
Southwestern Australia (Figure 7.1d)
Breeding seasons of birds in the southwestern region of the State
of Western Australia have attracted interest from a number of early work-
ers beeause of instanees of out-of-season breeding. Davies (1979) pro-
vided the most reeent and definitive account, including an analysis of five
seasons of monthly data eollected by Serventy between 1959 and 1964
from a Zebra Finch eolony breeding in the wheat-belt at Grass
Valley, York, 100 km east of Perth. This eolony was also studied by
Imme1mann for three to four months in the spring and summer of
1959-1960 (Immelmann 1962a). York is semiarid with a mean annual
rainfall of 445 mm, but preeipitation falls almost exclusively in winter
due to moisture-bearing southwesterlies from the lndian Ocean, so there
is a hot dry period of three to four months duration in the middle of sum-
mer. This asynehrony of maximum temperatures and maximum
precipitation results in a bimodal plant growth pattern with a
pronounced spring peak followed by a deep summer trough that rises to a
smaller peak in autumn (Nix 1976). The smaller autumn peak is not only
a response to the onset of early winter rains, but also to heavy episodic
rain caused by cyclonic aetivity in late summer. Serventy's pooled data
for Zebra Finehes breeding at York eorrespond well with this pattern.
There is a non-breeding period over the winter months followed by a
nine-month breeding sea son that peaks in October and November,
declines in the summer months, and peaks again slightly in April. Davies
couid find no dircet corre1ation between the number of breeding nests
started each month and rainfall figures for that month, nor couId he fine!
7 Breeding periodicity 123
any correlation with mean monthly temperatures, although mean mini-
mum monthly temperatures are prabably more relevant. Immelmann
(1962a, 1965a) was impressed by the sudden inhibitory effects of low
temperatures on the onset of spring breeding by Zebra Finches; 50 per
cent of clutches were abandoned during a cold snap in mid-September
1959. Unfortunately, no dietary studies were conducted during these sur-
veys to determine the availability of grass seeds, which may have been
responsible for much of the variation acrass seasons.
Arid Western Australia (Figure 7.1b,c)
Information on the timing of breeding by Zebra Finches in the arid
regions of Western Australia comes from two sources: scattered Nest
Record Scheme data across a vast area and an intensive three year study
conducted at one location by Davies (1977a). Data fram the former can
be pooled into two regions: 'Goldfields' and 'Pilbara'. The Goldfields
graph (Figure 7.1c) includes sites in the inland subdivision of the south-
west below 25°S, and the Pilbara graph (Figure 7.1b) includes locations
between 20° and 25°S. Mean annual rainfall is less than 250 mm in the
Goldfields and is irregular throughout although more seasonal in the
north (mainly occurring in autumn and early winter) than in the east and
south. In the Pilbara region, mean annual precipitation is greater and
more seasonal than Goldfields with most falling in late summer. The
graphs show that it is possible for Zebra Finches to breed in most
months of the year in both regions but there is a distinct peak in Septem-
ber and October. There is a smaller peak that occurs in early autumn in
the Goldfields and in late autumn-early winter in the Pilbara. Data for
the Pilbara are few and must be interpreted cautiously but they are con-
sistent with Carnaby's (1954) observations that birds in this regio n nest
predominantly in winter and spring but can nest at any time of the year
if there are sufficient rains.
In 1973, 1974 and 1975 Davies (1977a, 1979, 1986) monitored breed-
ing activity of Zebra Pinches for 25 continuous months at Mileura Sta-
tion. The mean annual rainfall for Mileura is 198 mm and occasional
heavy falls can occur in summer and winter but not every year (Davies
1986). He estimated the month of hatching by back-dating the age of
subadults from among 20-30 birds collected each month. The degree of
ossification of the skull was used to age birds up to 120 days post-hatch
(Serventy et al. 1967), and so pravided a means of determining when
breeding occurred up to four months previously. Although the number of
young hatched each month is difficult to extract fram Davies' (1977a)
Figure 1, my analysis shows that in 1973 hatching occurred in every
month with a slight peak in August and September. In 1974 hatching
occurred in ]anuary, November and December with a few hatching in
February, May and October. Data were also available for seven
non-continuous months in 1975 when breeding was backdated to April,
124 THE ZEBRA FINCH
May, June and August; none were estimated to have hatched in January,
February, March and July. Despite strong variation among years, the
data, taken together, indicate that Zebra Finches at Mileura have the
ability to breed in any month and in sorne years may breed continuously
at low levels, while in others there may be no breeding at all for seven
months or longer. However, they breed slightly more in spring and early
summer.
The relationship between rainfall and breeding at Mileura is not
straightforward, and my analysis of the data differs slightly from that of
Davies. Rainfall in winter 1973 was followed by a good burst of breeding
the following spring, but good rains in winter 1974 only resulted in weak
early summer breeding. The effects of rainfall on seed production is the
crucial factor. Davies (1977a) found three main species of annual grass
seeds (Aristida contorta, Eragrostis australis and Eriachne spp.) in the
crops and believes that the availability of plentiful supplies of these seeds
was responsible for the breeding of Zebra Pinches at Mileura. Insects
were rarely found and can be assumed to be of no importance in the dieL
J. Mott (1972), a plant ecologist, found that these species of grasses at
Mileura normally germinated after heavy falls of rain in summer and they
flower and set seed in autumn. They can also germinate after winter rain
and seed in spring if insufficient rain has fallen the previous summer, but
the rate of germination and the extent of seed set is much lower than
those produced by summer rains. Therefore, ripening seed should nor-
mally be available in abundance in late summer and early autumn in most
years but can also be found in the spring of sorne years. Thus, the spring
breeding in 1973 must have resulted from germination of seed in winter
caused by the heavy winter rains. Further rains in October and November
1973 probably resulted in the germination of more seed and extended the
breeding sea son into January and early February. The failure of the
March and April rains to produce a new burst of winter breeding is prob··
ablya consequence of poor germination after the good summer germina-
tion (Mott 1972). Subsequent breeding episodes in 1974-1975 appear to
be preceded by rainfall events with variable lag periods. The tightness of
the relationship between rainfall and breeding can best be established by
examining the crops of nestlings for seeds and relating these to the seed
set and germination of the grass species concerned.
Eastern Australia (Figure 7.1l,h)
The two arid regions in inland eastern Australia for which there are Nest
Record Scheme data show pronounced peaks in breeding during the
spring months. The spring peak occurs earlier in western Queensland
(Figure 7.11) than in western New South Wales (Figure 7.th), presum-
ably beca use of the more prolonged winter in the latter region. The
Queensland data are relatively few and a probably biased towards the
winter and spring months when the area is more frequently traversed by
7 Breeding periodicity 125
amateur data collectors, whereas the data for New South Wales are more
substantial and reliable since systematic samples were gathered by resi-
dents in the Broken Hill area. Rainfall is low (mean annual rainfall 244
mm) and aseasonal in distribution.
Table 7.1 Length of the breeding sea son expressed as the number of 'equally good
months for breeding' (EGM)a for five Zebra Finch populations for which at least 12
months continuous data were available
Alice 5prings Zann et al. (1995) (3) 23°45'5 133°52'E 7.6 10.6
Armidale Kikkawa (1980) (2) 30°30'5 151°40'E 7.0 8.9
York Davies (1979) (5) 31°52'5 116°47'E 6.9 8.0
Griffith Frith and Tilt (1959) (1) 34°00'5 146°0'E 6.6 8.6
Wunghnu Zann (1994a) (4) 36°09'5 145°26'E 6.5 7.6
a EGM is calculated from the formula, exp (-L Pi In Pi), where Pi is the proportion of nests started
in the ith month (MacArthur 1964). Values range from 1, when al! nests started in the same
month, to 12, when an equal number of nests started each month.
b Number of years of continuous monthly breeding samples.
C Based on Wynham's (1986) regression of length of breeding season at ten locations in the Aus-
tralian region.
126 THE ZEBRA FINCH
the five study sites where Zebra Finch breeding was monitored, the EGM
was greater than that predicted by Wyndham's line of best fit for the
Australian region. This means that Zebra Finches have longer breeding
seasons than most other Australían species from equivalent latitudes.
The breeding period for Alice Springs was longer than reported by
Wyndham for any location in the world.
In further analysis Wyndham (1986) failed to find any correlation
between mean annual rainfall and EGM in the Australian region, nor
any differences in EGM between Africa and Australian localities, and
concluded that irregular rainfall in Australia did not prolong the breed-
ing season in birds of the arid zone as predicted by Immelmann
(1963a,b). This was supported by the finding that in Western Australia
there was no difference in EGMs for a suite of species from coastal, more
mesic localities, and those from inland arid ones. In light of this, Ford
(1989), in his review, concluded that opportunistic breeding in Australia
was exaggerated and that seasonal breeding prevailed in the arid inland
as elsewhere. However, it must be pointed out that all Wyndham's local-
ities carne from southwestern Australia between latitudes 24° and 32°S,
and could in no way be considered representative of the whole arid zone,
consequently, his three conclusions must rema in tentative at this stage
until more representative data become available. The exceptionally long
breeding season of Zebra Finches at Alice Springs is more consistent with
Immelmann's hypothesis than that of Wyndham's. This may be related
to the fact that seasons are less predictable in central Australia than they
are in southern Western Australia (Nix 1976).
Opportunistic breeding
Although breeding seasonality is low in aH populations of Zebra Pinches
studied to date there is a consistent trend for stronger breeding in the
spring months. This does not mean it is an annual event, rather the pat-
tern only emerges in sorne locations when records are summarised over a
number of years. Moreover, in the five studies where breeding activity
was monitored each month for a number of successive years, it was clear
that no two years were alike, presumably, beca use of differing environ-
mental conditions. Taken together, these observations emphasise the
highly flexible nature of breeding in the Zebra Finch, with, nevertheless,
a predisposition for breeding in the spring months. The data, however,
are incomplete for the northern part of the range, and a different pattern
may emerge in this region with good sets of nesting data.
Before any long-term breeding data were available, Zebra Finches
were widely regarded as opportunistic breeders in arid parts of Australia.
Short-term and incidental observers in the arid zone were impressed by
the ability of Zebra Pinches and other arid species of birds to breed in
response to sporadic rainfall, no matter what the season, in order to
7 Breeding periodicity 127
exploit favourable conditions that arose shortly afterwards. Observers
included Carter (1889), Carnaby (1954), Serventy and Marshall (1957)
and Harrison and Colston (1969) in arid regions of Western Australia;
Keast and Marshall (1954) and Immelmann (1963a,b, 1965b) in central
Australia; and McGilp (1923) in the Lake Frome region of South Aus-
tralia. The effect of rainfall on breeding was especially impressive when
observers, such as Immelmann (see below) , witnessed birds starting to
court and nest-build while rain was still falling. Similarly, Carnaby
(1954) was also impressed by the dramatic impact of localised rainfall on
plant growth and breeding by birds in the Pilbara of Western Australia.
Here winter rains from locally restricted thunderstorms produced small
patches, or 'oases', about two by seven kilometres in one case, of green
lushness in which breeding activity was prolific, and outside of which
nests were far less common.
These incidental observations would indicate that breeding by Zebra
Pinches and other similar species always follows rainfall; it is implied
that rainfall at any time in the arid zone always produces a supply of
ripening grass seeds for the provisioning of young. However, our obser-
vations at Alice Springs (see below) and those by Maclean (1976) in
western New South Wales and Davies (1977a) in Western Australia
show that in certain circumstances there may be a delay between rainfall
and breeding, and the response may be weak. Naturally, the key factor is
whether rain always stimulates a growth cycle of grasses that results in
ripening seeds and whether Zebra Pinches always exploit these for breed-
ing every instance they are available. Growth and seeding of arid zone
grasses is complex and varies regionally according to climate, topogra-
phy and species composition.
Slatyer (1962) found that the timing of grass growth and seeding in
central Australia not only depends on the timing of rainfall and the pre-
vailing soil temperatures and available nutrients, but it also depends on
what rainfall has occurred in previous seasons. Normally, to pro mote
plant growth in this region a certain quantity of 'initial effective rain' is
needed to germinate annuals and to stimulate the regrowth of perennials,
but this mus! be followed by 'effective carryover rain' to continue vegeta-
ti ve growth to a stage where flowering and seed set can be achieved.
Slatyer (1962) found that the prevailing temperature, through its effects
on evaporation rates, determined what quantity of rain would be effec-
tive: 7-16 mm of initial rain was needed in winter and 22-29 mm in
summer; carryover rains needed to be double these amounts. A single
large faH of rain could sustain moisture levels so that it could serve both
to initiate and carry over a cycle of growth. Another important factor is
that the seeds of different species of plants germinate at different temper-
atures. In general, the seeds of most grasses are sensitive to frosts and
remain dormant during the frost period although there are sorne short-
lived annuals (Five-minute grass Tripogon loliiformis, Armgrasses
128 THE ZEBRA FINCH
Brachiaria muliiformis and B. gilesi, and sorne species of the Nineawns
Enneapogon spp.) that will respond to effective winter rains, but most
species will not germinate or re-grow without the warm rains of summer.
In contrast, herbs, mostly daisies, germinate and grow when effective
rainfall occurs in winter. 1 encountered this phenomenon in 1986 when 1
stayed at Alice Springs for eight months. In May 1986 we began moni-
toring the breeding activity of Zebra Finches towards the end of a pro--
longed dry period in which only 143 mm of rain had been recorded for
the previous 17 months; all grass was dead and consisted mostly of over-
grazed 'butts' (Zann el al. 1995). Hungry Zebra Finches were digging
into the soil for scarce seeds and large flocks were coming to the few
remaining drinking sites that still held water. Although many nests were
checked, all were empty, and no subadults were seen among the large
flocks inspected through binoculars or captured in mistnets-indicating
that no breeding had occurred for at least three to four months previ-
ously. Between 19-23 June, 25 mm of rain fell and another 151 mm feH
a week latero Within a few weeks this exceptional winter rain caused the
rapid germination and growth of forbs and soon resulted in the
renowned 'carpet of daisies' across the landscape. The exotic perennial,
Buffel Grass Cenchrus ciliaris, the dominant species of grass in the dis-
trict and the mainstay of local Zebra Finches, did not begin to regrow
until early September, and seeds did not ripen until early October. Not
surprisingly, Zebra Finches did not breed with the rain but waited until
the food supply for their young was available. The first eggs were found
on 21 August and the first young had green C. ciliaris seeds in their crops
in mid-September (Zann el al. 1995). Thus, the Zebra Finches had found
green seeds several weeks before we did and displayed the typical spring
bout of breeding although not in the frenzied burst that we had expected
from descriptions of earlier authors. This was probably a consequence of
the fact that the grasses were not growing with the expected vigour and
abundance despite heavy falls of follow up 'grass rains' in the warm
months of September, October and November. Local plant ecologists had
never experienced the situation where heavy winter rains were followed
by heavy summer rains. The poor growth of perennial grasses in summer
was thought to be due to the fact that the heavy layer of forbs produced
by the winter rains had locked up the necessary soil nutrients and their
shade perhaps prevented or retarded the germination of annual grasses.
Despite our experiences in 1986, subsequent monitoring of Zebra
Finches at Alice Springs showed that not only was winter breeding possi-
ble, but it could be prolific. We believe that heavy rains in March 1988
were responsible for the strong breeding activity recorded for the follow-
ing June, JuIy and August. Similarly, in the following year, another burst
of breeding in May, June and July appeared to be a consequence of
heavy rain in the previous March (Figure 7.2). Unfortunately, the seeding
phenology of thc grasses was not monitored after 1986 so nothing can
7 Breeding periodicity 129
be conduded about the availability of the of grass seeds that, no doubt,
made breeding possible. One must condude therefore, that in the central
arid zone not only is rainfall spasmodic and unpredictable, but its effects
on the growth of grass and the subsequent production of seeds are also
fairly unpredictable. In order to exploit every available opportunity for
breeding Zebra Finches must track the occurrence of ripening grass seeds
in both time and space and respond to any environmental eues that
might predict the impending production of this resource.
Summary
Nothing is known about the breeding periodicity of the Lesser Sundas
Zebra Finch, but it is expected to breed towards the end of the monsoon
before the dry sea son sets in. The breeding period of Australian Zebra
Finches is long and highly flexible. The minimum duration is eight
months in southem latitudes, but may extend to fifteen months in central
Australia in sorne seasons. The breeding period is longer than that of
other Australian species of birds and increases in duration from high lati--
tudes to low latitudes. Breeding can occur in any month of the year in
sorne populations, but most show sorne seasonality with a winter lull
and a spring peak, and sorne also show a second peak in autumn. North-
em populations have a summer peak rather than a spring peak but more
data are needed from this region. Breeding in central Australia is excep--
tional-although there is a spring peak, breeding is non-seasonal and a
whole year may pass without any breeding attempts; when suitable con-
ditions arise breeding will begin whatever the month. Availability of
maximum supplies of ripening grass seeds is the critical factor determin-
ing when breeding will occur and this depends on the plant growth cycle
for the particular climatic region. This is fairly seasonal in all locations
except central Australia, where Zebra Finches are opportunistic breeders
in the sense that they exploit any production of ripening grass seeds irre-
spective of the time of the yeaL Here the timing of grass growth and seed
production is complex but mainly depends on when good falls of rain
occur; however, this is highly erratic in space and time. A recent study in
central Australia did not verify the renowned observations of Immel-
mann in which Zebra Finches made an immediate breeding response to
rainfall, rather, there was a lag of two to three months between the first
faH of rain and the first eggs. The proximate factors that stimulate breed--
ing in Zebra Finches probably vary within and among climatic regions
and are subject to strong selection pressures.
Except for the populations from central Australia, low tempera tu res
inhibit breeding activity in Zebra Finches. Temperature thresholds for
inhibition and disinhibition of breeding probably vary in different parts
of the range. Dehydration inhibits spermatogenesis and ovulation.
8 Populations
'At the spring near Mt Ultim the Chestnut-eared Finches were present in countless
myriads, rising in clouds with a noise like a rushing wind from every shrub and
tree within a quarter-mile of the water, clothing the leafless branches of dead
trees with their quivering forms looking in the distance like grey butterflies as
they sought a resting place, and covering the rocks round the spring itself with
their droppings till these at first sight and to the uncritical observation looked as
if a light mantle of snowflakes partly hid their texture from view.'
J. B. Cleland in central Australia, 1931.
600~1--------------------------------------' 350
500 I 11 300
]; :g, 250
§ 400 I I I §
8
1i 200
1iE 300 I lf
El e~ 150
200 e
1 1'I
~ 100
::;;
100
~ !" 50
o
D J D J J D J D o J 1974 1975 1976 1977 1978 1979 1980 1981
1968 1969 1970 1971 1972 1973 1974
350 160
I ,
,.,300 140
'"
;: 250
~
\ "O
120
\ ~ 100
~ 200
" \ ~a; 80
~150
1iE 100 -\/\. ~ 60
Z"
~ 50 V "'- .A>- ... 40
20
o o
1986 1987 1988 1989 1990 1991 1992 1985 1986 1987 1988 1989 1990 1991 1992 1993
Fig.8.1 Yearly ehanges in abundanee of Australian Zebra Finehes from four loealities. (a) Mean number of Zebra Finehes eensused
pe! watering point surveyed at Erldunda (25°S, 135°E), 200 km southwest of Aliee Springs (modified from Corbett and Newsome
1987); (b) mean number of Zebra Finehes mistnetted per day eaeh Oetober at Mileura station, Western Australia (from Table 4 in
Davies (1986)); (e) number trapped by walk-in trap set for ane day eaeh August at Aíiee Springs (Zann et al. 1995); (d) number
tranned hv walk-in tran eaeh AUlwst at the Danaher colonv. northern Victoria (this studvJ.
8 Populations 141
apparent crash in numbers after 1973 is an artefact of the method of
census at water-holes-good rains fell at this time so birds no longer
carne to the permanent watering points surveyed. Large fluctuations in
numbers of birds trapped at a watering point also occurred at Mileura
(Figure 8.1b). This was a direct consequence of the amount of rainfall
that had fallen in the previous 12 months (Pearson correlation between
rainfall and numbers captured r = 0.92, n = 8, P = 0.001: S. J. J. F.
Davies, pers. comm.). Fluctuations at Alice Springs were also a conse-
quence of rainfall the previous year (Figure 8.1c; r = 0.84, n = 7, P =
0.02), but rain had no effect on abundance of wintering birds at the
Danaher colony (r = -0.18, n = 8, P = 0.67), since this is an irrigated
are a (Figure 8.1d).
Two attempts have been made to estimate the size of Zebra Finch pop-
ulations exploiting localised resources. At Mileura, Western Australia,
Davies (1986) banded and released all Zebra Finches coming to a water
point, 'Jindi Jindi', for a period of several days each October for eight
years. He captured 2,867 individuals over the survey period, including
1,398 birds caught over a four-day period in 1981, but recaptures were
quite low (total 120). He used the simple Petersen method to estimate
the number of birds coming into drink by comparing the proportion of
birds recaptured on the second day with that banded and released on the
first day. After he banded at Jindi Jindi in 1981 he checked six adjacent
water points for recaptures. He found none among 600 captured so
assumed that there was no immigration or emigration over the few days
trapping occurred at Jindi Jindi. His estimates of the population ranged
from 686 in 1976 (137 banded in three days) to 22,258 in 1980 (351
banded in two days). While not much confidence can be placed in the
accuracy of the Petersen method, since it does not produce standard
errors, these values give sorne idea of the order of magnitude of fluctua-
tions that can occur in natural populations.
In 1986 we also used capture-mark-recapture methods to estimate the
number of Zebra Finches visiting a walk-in trap baited with seed at Alice
Springs. Trapping data for four months satisfied the criteria for estimat-
ing population size using the Jolly-Seber method, which gives standard
errors (Caughley 1977). Values ranged from 320 ± 130 in August to 198
± 78 in November (Zann et al. 1995).
The size of Zebra Finch populations in arid Australia clearly undergo
wild fluctuations in response to annual rainfal1. Other aspects of the
population biology of Zebra Finches were investigated using capture-
mark-recapture methods at my study colonies in northern Victoria and
at Alice Springs.
142 THE ZEBRA FINCH
Survivorship
Banded nestlings
Survivorship of Zebra Finches banded at the Danaher colony was very
low (Figure 8.2)-67% were lost between banding age and day 35, the
age of nutritional independence, and this accounts for the initial steep
decline in survivorship. Most losses probably occurred after young had
fledged, hence the 15-20 day interval between fledging and nutritional
independence appears the most vulnerable part of the life cycle at this
colony. Unpublished observations of fledglings at Danaher are consistent
with this level of losses: by day 35 only 6 out of 44 young that fledged
were observed around the feeder at the walk-in trap and 9 out of 37
fledglings were observed at the roosting nests (Chapter 9). Most losses
during this interval are probably mortalities rather than movements from
the colony beca use extended flight of newly fledged young appears fairly
limited, although this should not be underestimated. Predators are the
likely cause of losses beca use food was supplied in abundance at the
walk-in trap, but it is possible that parents may have neglected the young
for sorne reason, perhaps when re-nesting. Losses after day 35 kept fairly
constant and remained so until they suddenly dropped again after day
540. These losses arise from mortalities and emigration. By day 35 the
improved flying ability enable young to avoid predators, and their com-
petence in dehusking seed may approach levels of efficiency found in
adults (Chapter 9).
--_._-----,._--------------
day 35- nutritional independence
14 50 1 00 160 220 280 340 400 460 520 580 640 700 760 820 880
Days post hatch
Fig. 8.2 Log-normal survivorship curve (proportion recaptured by days after band ..
ing) for 850 Zebra Finches banded in the nest at the Danaher colony 12-14 days
after hatching, based on known age at disappearance. The study encompassed six
breeding seasons 1985-1991. (After Zann and Runciman 1994.)
8 Populations 143
Only 23 % of nestlings banded reached 90 days of age, the median age
of first reproduction for birds breeding in their season of hatching. Just
65% of hatchlings reached banding age in the first place, consequently
the probability that a hatchling will survive to breeding age was only
0.15 (0.23 x 0.65). Mean annual survivorship for the first 12 months of
adult life was four per cent and the estimated life expectancy at hatching
was 51 days.
Losses of Zebra Finches at Danaher between fledging and age of first
breeding are extremely high, especially when one considers that such a
short span of time is involved (about 70 days). Nevertheless, the value
(77% losses) is within the range found for other species of birds
(42-86%; Newton 1989), but few species reach sexual maturity as early
as Zebra Finches. Data on survivorship of nestlings to age of first breed-
ing are needed from Zebra Finch colonies in other parts of the range in
order to place these findings from Danaher in perspective.
At Danaher, the proportion of males and females surviving from one
month to the next (age-specific survival) was not significantly different.
There was a trend for females at both Danaher and Padgett colonies to
live slightly longer than males although this just failed to reach statistical
significance (Table 8.1). The oldest known-age bird was a female that
disappeared when four years and five months of age although a female at
the Padgett colony disappeared five years and three months after first
capture as a mature adulto The longest span between first and last cap-
ture at Alice Springs was 13 months for several birds caught as mature
adults. At Mileura, Davies (1986) recaptured one bird three years after
banding. Therefore, although the vast majority of individuals appear to
have very short lives, sorne can live up to five years, or more, in the wild.
The potentiallifespan of domesticated Zebra Finches held under optimal
breeding conditions is 5-7 years (Burley 1985a).
Table 8.1 Estimates of lifespan based on the interval in days between first and last
capture at walk-in traps for Zebra Finches hatched in the Danaher and Padgett
colonies that reached at least 35 days of age a
Danaher Padgett
a Data include young banded in the nest and those caught before 35 days. There were no significant
differences between the sexes (Danaher: Wilcoxon two-sample Test Z = -1.88, P = 0.06 and
Padgett: Z = 1.86, P = 0.06). (From Zann and Runciman 1994.)
144 THE ZEBRA FINCH
Table 8.2 Population parameters for three colonies of Zebra Pinches in northern
Victoria and one colony in Alice Springs; estimates are derived from the capture,
mark and recapture of mature adults (¿ 100 days of age at first capture) at baitcd
walk-in traps set at monthly intervals
--_._._
.. _---_ .. _ - _ . _ - - - - - - -
Colony Number trapped Per cent recaptured Annual mortality (%) SlO (years)'
one month after
first tra pping
._-_._---------
Danaher" 68"7 59 82 1.2
Padgen a 185 53 n 1.5
Cloverlea" 268 61 n 1.6
Alice Springs lb 613 18 96 0.1
Alice Springs nb 454 31 83 0."1
----------_.-
a Zann and Runciman (1994).
b Period I (June 1986 to September 198"7, Period II (July 1988 to August 1989); from Zann et al.
(1995).
, Interval after first banding when 10% of birds could still be recaptured.
8 Populations 145
~ Padgett 1981-83
Ol
e
";;;
"2: 0.1
::::l
fJ)
e
O
'eo
o.
2
o..
0.01 L - _ - - L_ _- ' -_ _~_+_-----<~---1.----'-----1
O 2 4 6 8 10 12
Months
those found here for the Zebra Finch (Woinarski and Tidemann 1992).
Nevertheless, in a pilot study, Tidemann (1987) found that the Zebra
Finch had the lowest recapture rate among seven species of estrildines
trapped over a four day interval at drinking points in the Top End.
Low survivorships estimated from capture-mark-recapture studies do
not necessarily mean that mortalities are extreme because 'dispersal'
could also make birds unavailable for recapture. Trap shyness, which
also reduces rates of recapture, is more significant with mistnetting than
with walk-in traps (Zann and Runciman 1994). Unfortunately, in cap-
ture-mark-recapture studies, there is no easy way to distinguish effects of
mortality from those of 'dispersal', but a number of lines of evidence
suggest that 'dispersal' may be a significant factor accounting for the
high losses of banded Zebra Pinches from the trapping sites.
Sex ratios
Sex ratios of the Lesser Sundas Zebra Finches are unknown, however,
secondary and tertiary sex ratios have been examined at a number of
widely separated populations in the Australian subspecies (Burley et al.
1989). Secondary sex ratios are those found in young at the end of
146 THE ZEBRA FINCH
parental care and tertiary ratios are those found at adulthood.
Tertiary sex ratios were slightly male biased, but not significantly so,
acrass six populations (Table 8.3). This trend was consistent (test of
homogeneity: G 1 == 0.53, P == 0.97), and when po oled the sex ratio for
10,741 adults was 52.0% males which was significantly different fram
parity (G == 8.8, P :::: 0.003). There was no significant between-year vari-
l
ation in the tertiary sex ratio at the northern Victorian colonies nor did
method of trapping (mistnetting vs. walk-in trap) affect the sex ratio
(Burley et al. 1989).
Reliable data for calculating secondary sex ratios were obtained fram
Padgett and Danaher where young were sexed between 35 and 40 days
of age, just after they have become nutritionally independent of their
parents. Whereas, Padgett birds (n := 148) were 52% male (Burley et al.
1989), Danaher birds (n == 407) were 44% male (Zann and Runciman
1994); however, the difference was not significant (G == 2.8, P == 0.7) nor
l
m m m m m f m
Adults 223 209 402 381 372 315 1908 1783 308 290 1468 1334
Percentage male 52 51 54 51 51 54
Young surviving 29 18 110 96 121 125 11 6
to adulthood
Percentage male 62 53 49 65
Total adults 252 227 512 477 493 440 2555 2385 308 290 1468 1334
Percentage male 53 52 53 52 51 52
(e)
---~2-
0-· /~oº, f
···~
."?,
.-.....
long drought
Fig. 8.4 Hypothetieal sehemata illustrating the strueturing of populations and the
types of movements made by Zebra Finehes during favourable eonditions (a), during
dry seasons (b), and during prolonged drought (e). The large eircles represent the
home range of populations whieh are eomposed of separate eolonies (XXX). The
arrows show movements of individuals; arrows within the home ranges represent
'extended local exeursions' among eolonies; 'extra-home range wanderings' are
shown by the arrows that leave the home ranges; 'large-seale movements' are repre-
sented by the long arrows in (e). The small hatehed eircles represent watering points
and the blank ones represent watering points that have dried up.
The proximate factors that cause individuals to move from the nesting
colonies are not known, but it is reasonable to assume that the guest for
the key resources of seed and water, which are unpredictable in both
time and space over most of the distribution, is the principal factor
responsible. Increasing densities of predators and parasites may also
cause departures from colonics. If movements are density dependent then
8 Populations 155
the high levels of transients from study colonies where seed is provided
ad libitum at walk-in traps require explanation. Why should birds leave
sites with abundant, high-quality food, abundant nesting sites and abun-
dant water? Conceivably, new arrivals at a study colony would follow
locals to feed at the seed hopper in the trap and seeing only a small cirele
of seed surrounded by dozens of finches feeding shoulder to shoulder,
may underestimate the sustainability of the resource and soon move off
to another part of the home range. AIso, the renewal capacity of the seed
at the hopper can only be assessed from several visits and sorne newcom-
ers may not stay long enough to make them. This hypothesis could be
tested experimentally. It still does not explain why residents and birds
hatched in the colony make excursions elsewhere when food, nesting
sites and water are available ad libitum. There is no direct evidence of
interference competition forcing sorne individuals to move away. Envi-
ronmental factors may not be solely responsible for forcing birds to
move to other colonies; sorne movement may be voluntary. There is
sorne suggestion for this in recapture data at Danaher where there were
sex differences in recapture rates between young males and females. Emi-
gration by males appears to be affected by the local abundance of food
whereas female emigration is not, and may be intrinsic, that is, under
more direct gene tic control. This is the subject of a recent review of dis-
persal in mammals by Lidicker and Stenseth (1992).
Summary
Size of populations fluctuates sharply from one year to the next, and
there is a strong positive correlation between numbers and rainfall in the
arid zone, but not in irrigated country. Banding studies reveal that Zebra
Finches have very high rates of mortality (67%) between fledging and
nutritional independence. The maximum life span is five years and
females may live slightly longer than males. Survivorship rates after day
35 are extremely low, but losses due to mortality are confounded with
losses due to movements away from the study sites. Immigration rates
are very high; only 22 % of all adults captured at the Danaher colony
hatched there and only 23.5% of breeding adults were birds of the
colony, the rest were from other colonies. Among breeding birds there
was no sex-biased philopatry in those hatched at the Danaher colony nor
was there any sex bias among immigrants. The sex ratio at adulthood
(tertiary sex ratio) is slightly male-biased (52% males overall) and the
sex ratio at 35-40 days of age (secondary sex ratio) is slightly female-
biased. The secondary sex ratio may depend on the seasonal abundance
of food. Emigration of young males may depend on availability of food
at the colony but that of females is independent of food supply. The pat-
tern of changes in the sex ratio s are not elear and more data are
required. Indirect evidence from a number of sources indicates that even
156 THE ZEBRA FINCH
when seasonal conditions are good, Zebra Finches are highly mobile.
They traveI continuously among the nesting colonies over a very large
home range. More-extensive nomadic movements beyond the home
range may occur under sorne conditions. A population consists of a num--
ber of different flocks that nest in discrete colonies, but reIy on the same
source of permanent water during unfavourable seasons. When this
so urce dries up the surviving members of the population are forced to
take more extensive movements beyond the home range in order to find
new so urce s of water, and food, and birds appear in locations where
Zebra Finches have never been seen previously. Although they are not
included among the long-distance dispersers of inland Australia, Zebra
Finches are extremeIy efficient dispersers, and have reached most of the
artificial supplies of surface water created since European settlement.
9 Social and reproductive behaviour
'The courtship of the male Zebra Finch consists of auditory, static-visual, and
dynamic-visual elements, whilst that of the female is primarily static-visual.'
D. Morris 1954.
Social behaviour
Flying and flocking
There are several modes of flight. When travelling long distances, such as
to and from food or water, flight is strong and determined. Birds quickly
climb to an altitude of 10-20 m and fly rapidly in a straight direction
with little undulation. They can be followed through binoculars for no
more than about 60 seconds before the specks disappear into the dis-
tance, at approximately 0.5-1 km away. Pairs usually fly on the same
level, about 1 m apart, one bird leading. When flying short distances,
such as those between different parts of a colony, they fly more slowly,
with greater undulations, at altitudes between 0.5 m and 2.5 m. Should
they be surprised in flight by a predator, such as a faleon, they will
descend to about 0.5 m and fly in a rapid zig-zag course to the nearest
thick bush, where they shelter in its centre.
Wind tunnel studies show that Zebra Finch flight is geared for distance
rather than speed. When the wings are folded between wing beats,
maximum lift and gliding distance is obtained at the expense of speed
(Csicsáky 1977).
Feeding and drinking flocks are reasonably tight and fairly well synchro-
ni sed (Chapter 4). Coordination among flock members appears to depend
mainly on vocalisations, and their cohesion and orientation in flight is
probably helped by the conspicuous white tail rump. Zebra Finches flock
158 THE ZEBRA FINCH
throughout the year, but size of flocks depends on enviranmental condi-
tions and whether birds are breeding or not (Chapters 4 and 8).
Clumping and allopreening
Zebra Finches are 'contact' species. During resting periods at any time of
the year, sexual partners may be found perched or squatting in contacto
Frequently, dumping is interrupted by auto- and allopreening. There are
no special allopreening invitation postures other than erecting feathers
on the head and throat and exposing them to the partner (Figure 9.1). In
free-living flocks, members of pairs are usually identified as those that
dump with, and allopreen, one another. Parents allopreen and dump
with offspring, and siblings will also dump together and allopreen, but
this ceases after 40 days of age when the male sexually dimorphic
plumage begins to appear. Whereas unmated adult females may dump
together, males do not and if their individual distance is infringed they
move apart or bill-fence. Low temperatures do not increase the fre-
quency of dumping, but sick Zebra Finches will dump more than
healthy ones.
Domesticated adults are not as discriminating as wild birds and males
may sit in contact and even allopreen. Males of white morphs, which are
sexually monochramatic, dump together and wild-type males are not
inhibited fram dumping with white males (Immelmann 1959). This
shows that the individual distance between males is a function of male
plumage and that vocalisations and other behaviour have little effect. In
a series of experiments, Immelmann (1959) showed that colour of
plumage of a conspecific releases the dumping response, but the method
and orientation of dumping depends on the red bill. AIso, a model in a
low squatting posture is preferred over one in a less squatting posture
Fig. 9.2 A female uses bill-fencing to reject the unwanted advances of a maleo
160 THE ZEBRA FINCH
breeding nests, and the perches used to approach nest entrances. A paír
fights as a team and can easily dominate superior numbers of single
birds, for example, when fighting over a roosting nest. The dominance of
a male can change depending on whether his female is present or noto
The appearance of females wiU cause outbreaks of fighting among hith-
erto peaceful males (Immelmann 1959). Both males and females defend
their partners against approaches by same sex-rivals, and 1 have seen
females supplant single females courted by their partners. Immelmann
(1962a) states that males defend food against every one, including the
partner, but 1 have not seen this even among starving birds in central
Australia. Nest owners, especially incubating females, wiU defend the
nest strongly against strange pairs that might inspect the nest or come
too close; long chases over distances up to 100 m may ensue. Neighbours
are tolerated if they do not come too close. Copulating males also elicit
supplanting attacks from nearby males (and sometimes females) who
knock them off the females' backs. Females wiU use supplantings to
reject the advances of courting males. If small passerines come too close
to the nest, Zebra Finches will also supplant, and sometimes chase.
These include competitors, such as House and Tree Sparrows (Passer
domesticus and P. montanus, respectively), and nest predators, such as
Brown Honeyeaters Lichmera indistincta; however, harmless species,
such as thornbills (Acanthisa spp.), are ignored.
No dominance orders are evident in free-living colour banded birds,
but in captivity sorne individuals are consistently more aggressive than
others when held in unisex groups; they initiate more fights and win
more, and a rough linear order may be formed, but this changes daily
(Evans 1970; Ratcliffe and Boag 1987). The most dominant bird usually
wins control of the highest nest box in the aviary. However, there are
many retaliatory attacks from inferiors on dominants so that no linear
orders form, furthermore inferiors may win over dominants when
defending their own nest boxes. In domesticated Zebra Finches, aggres-
sion in all-female groups is not as freguent as that in all-male groups
(Butterfield 1970). In both domesticated and wild-caught birds, homo-
sexual bonds may form in unisex groups. Partners clump, allopreen and
cooperate as a team in the defence of resources and may become domi-
nant in an aviary. If reunited with their heterosexual partners there is an
increase in the level of aggrcssion cven when density is held constant
(Butterfield 1970).
In a series of carefully planned experiments, Caryl (1975) found that
domesticated mate Zebra Finches could be provoked to fight at the sight
(but not the sound) of a female and the amount of fighting depended on
how close the female was and whether the male was paired with her or
noto Sexually provoked aggression cannot be entirely explained by com-
petitíon among males for females, or defence of the sexual partner.
During pair formation Clayton (1990a) found that aviary-bred
9 Social and reproductive behaviour 161
Australian male Zebra Pinches were more aggressive than their smaller
counterparts from the Lesser Sundas.
Reproductive behaviour
Pair bond
Single birds of the Australian subspecies are found in most flocks
throughout the year and here they meet potential partners. Single birds
may be youngsters looking for their first mate or adults who have lost
theirs. On a population basis, there was no significant departure from
parity in the tertiary sex ratio (Chapter 8). However in 48 winter trap-
ping sessions at the Danaher colony, adult males were in excess of
females in 31 months, and females in excess of males in 15 months. Sur-
pluses could be quite large, for example, in May 1986 there were 22
adult males and only 8 adult females, whereas in August 1989 there were
23 males and 36 females. Mortalities within the colony and high levels of
immigration mean that pairing opportunities and levels of competition
for partners are quite variable from one month to the next.
Bond formation
There are no special pair-formation displays. In the wild, single males,
either mated or unmated, will approach any single female to initiate a
sexual encounter. Sorne perform exaggerated greetings and begin the
stage 1 courtship waltz (see later), others simply confront the female and
direct a number of song phrases at her, and sorne mated males omit all
courtship and simply jump straight on the female's back without prelimi-
naries. If the female is already paired, but her mate is incubating or else-
where, she may ignore him or supplant him. Uninterested females flee,
but if interested they will respond with a Head Tail Twist greeting (see
later) of their own and may even begin the stage 1 dance. Occasionally,
single females will take the initiative and greet single males, but mated
females have not been observed doing so. If both birds are receptive,
courtship does not proceed very far, but gradually over the course of the
day, or days, both birds spend more time together and courtship
becomes more frequent, although it still does not lead to copulation. The
first sign that a bond has formed is when the pair sit in contact and allo-
preen and this is the criterio n chosen in most studies. Copulation tends
to occur sometime after this when the pair begin to nest-build. There-
after the partners are inseparable unless incubating or brooding and their
activities are synchronised. Nest-building displays are not essential for
pair formation (d. Morris 1954) which can occur in both wild and
domesticated birds in the absence of any nest site or nesting material.
In the wild sorne bonds form within a day, although Immelmann
(1962a) found that sorne took many days. In unpublished experiments
162 THE ZEBRA FINCH
conducted at the Danaher colony during the non-breeding season
we removed mates of mated pairs and found that replacement pairing
usually occurred within several days. It appears that single birds in the
wild will form bonds when and wherever possible, irrespective of season.
Pair formation also occurs quickly in captive Zebra Finches of both
subspecies (Immelmann et al. 1978; Clayton 1990a). Using slightly dif-
ferent experimental designs, Caryl (1976) and Silcox and Evans (1982)
focused on the changes in behaviour that occur over the course of paír
formation in the Australian subspecies and arrived at almost identical
condusions. In each study, a pairing male was observed to fight more
with females with whom he did not paír, but sang more directed song
and copulated more with the female with whom he paired. He also
stayed doser to her, dumped with her more and synchronised feeding
and preening more. Most differences in behaviour between pairing and
non-pairing individuals reached significance on the second day of the
experiment, but dumping was not significant until day 4 or 5 in Caryl's
study, and day 2 or 3 in Silcox and Evans' study. Preferences fo1' a
potential mate can form soon after meeting in both subspecies
(Clayton 1990a); this was evident in most pairs in Silcox and Evans'
study within 30 minutes of meeting. Dunn (1981) also found that most
bonds were formed by the second day, with sorne forming as early as
20 minutes after introduction, whereas others took four or five days.
Dunn conduded that dumping was a means of strengthening a bond
that had already formed 1'ather than part of the formatíon process itself.
Potential partners cannot pair unless they can make tactile contacto
Visual and auditory contact and auditory contact alone were insufficient
for a pair bond to form (Silcox and Evans 1982; Clayton 1990a).
In captivity, incompatibilities often exist among potential sexual part--
ners and about 30% fail to establish bonds within about 10 days of con-
tact even when the sex ratio is 50% (e.g. 7/26, CaryI 1975; 6/21, Dunn
1981; 3/10, Silcox and Evans 1982). Ratdiffe and Boag (1987) found
that after 15 days, 10/36 females failed to pair despite a sex ratio of
66.6% males. By contrast, in two Bielefeld studies that used large
aviaries to examine the effects of cross-fostering on assortative mating,
the rate of pair formation was much higher: 95% (n = 64) of individuals
of both grey and white morphs paired within three days (Immelmann et
al. 1978) and 100% (n = 70) of available individual s of both subspecies
of Zebra Finches paired within an unspecific time (Clayton 1990b). Pos-
sibly, large aviaries may be a significant factor in the rapid formation of
pairs, since they allow the female space to escape overbearing males and
opportunities to initiate their own preferences.
Who chooses?
On the basis of his obse1'vations in the wild, Immelmann (1962a) con-
duded that the ultima te decision to form a pair bond or not, rests with
9 Social and reproductive behaviour 163
the female, and not the male-she may reject his proposal or accept, but
evidence for this conclusion is not given. On certain occasions a surplus
of single females will mean that they do not have a choice of males. It is
difficult to determine who is selecting whom, since the process is a
dynamic and reciprocal one. Male courtship is more active and conspicu-
ous than female courtship so when Dunn (1981) found that males selec-
tively courted the female with whom they eventually paired on first
meeting, it was not possible to tell if the male had selected the female for
her attributes or for the fact that she has selected him and encouraged
him in subtle ways. However, by means of a multiple-choice arena (the
'Finkodrome') that allowed the test bird to make a simultaneous choice
of 10 isolated stimulus birds, Clayton (1990a) found that both males and
females quickly discriminated among potential partners. Moreover,
within each sex there was agreement on which were the most desirable
partners. If the preferred individual was available later in the aviary pair-
ing occurred, but females were more successful in pairing with their ini-
tial preference than were males. This suggests that females have more of
the final say in pair formation than males. In a second experiment Clay-
ton (1990b) controlled the behaviour of the stimulus birds by placing
them behind one-way glass so that they could not respond behaviourally
to the test bird on approach and so influence its choice. When prefer-
ences tested this way were compared with those in subsequent tests in
which test and stimulus birds could interact freely through a wire screen,
there were significant changes in preference. Thus, for both males and
females the initial choice is based on the appearance of the potential
partner and the final choice is an outcome of a dynamic and reciprocal
interaction between them.
Females prefer males of their own subspecies and discriminate against
males that have been cross-fostered to the other subspecies (Clayton
1990b). Despite the tendency for the female to make the final decision at
pair formation, Clayton (1990b) found that her initial preference could
be overridden by that of the male if she was cross-fostered and he was
noto Possibly, normally reared individuals are less ambiguous about their
sexual preferences than cross-fostered ones, and their more persistent
courtship prevails over other shortcomings.
Levels of sexual incompatibilities are impossible to establish in the
wild, but given the tendency of males to court females indiscriminately, it
is likely that the female is the more discriminating sex, although failure
to form a pair bond during the breeding season could lead to lost breed-
ing opportunities. In both free-living and domesticated birds, two males
may compete for the same female. In domesticated Zebra Finches, both
Butterfield (1970) and Silcox and Evans (1982) found that male-male
aggression was very high at the start of pair formation, and Butterfield
(1970) and Ratcliff and Boag (1987) found that only the dominant males
formed pair bonds. Female-female aggression also occurs, but at a lower
164 THE ZEBRA FINCH
level than that of males and mostly after the bond is formed. Perhaps
males compete for access to the females, and the females fight to protect
the bonds that had started to formo
Pair formation in the Double-barred Finch is different from that of the
Zebra Finch. According to Immelmann (1962a, 1965a), there are no spe-
cial displays involved, but pairs form gradually at the onset of breeding
by a 'negative process' in which the two slowly restrict their dumping
and allopreening to the partner. In the non-breeding season, the pair
bond weakens in this species and partners will dump and allopreen with
other members of the flock once again.
Pair bond and pair formation in the three sexually monochromatic
species of Poephila ha ve been studied in controlled aviary experiments
(Zann 1977). Their bond is as tight as that of the Zebra Finch, or
tighter, and displays at pair formation are similar except that nest site
displays are an integral part of the process. Although both sexes also
show rapid preferences on first meeting, formation of the bond is slower
than with the Zebra Finch. This is due to mainly to high levels of com-
petitiveness in both sexes, and an 'unwillingness' to modify initial prefer-
ences so that reciprocal and mutual bond s can be established between
partners. Consequently, depending on the species, between 35-100% of
birds failed to form pair bonds over a 14-day periodo By contrast, Zebra
Finches appear more willing to pair with a non-preferred individual than
not to pair at aH. Conceivably, lives of Zebra Finches are shorter than
those of the Poephila and a partner, even an inferior one, is better than
none aH, since a pair needs to breed at the earliest and slightest opportu--
nity, whereas breeding in the Poephila is les s urgent since its timing is
seasonal. Possibly, extra-pair copulations (see below) are a more viable
option for the Zebra Finch and could offset the disadvantages of an infe--
rior partner.
Bond maintenance
Free-living Zebra Finches have a strong permanent pair bond which they
service constantly. They are an inseparable dyad, with the male usually
leading the female around the colony and in the flocks, except at breed-
ing when the roles are reversed (Birkhead et al. 1988a). When active, the
partners keep in contact and synchronise movements with Distance
Calls, Tets and Stack Calls (Chapter 10). Most resting moments are
spent in contact al1opreening, and potential sexual rivals are driven away
from the partner. At the start of nest building, nest ceremonies are per·-
formed, which Immelmann (1962a) believes serve to strengthen the bond
before a breeding attempt.
The paír bond in domesticated Zebra Finches is also strong and per-
manent, but divorces can occur (Morris 1954). If partners are physically
separated but allowed visual and auditory contact, or just auditory con-
tact, they are able to maintain the bond through the exchange of Dis-
9 Social and reproductive behaviour 165
tance Calls even when caged with a member of the opposite sexo If held
without any auditory and visual contact with their partners the bond is
broken and there is almost immediate re-pairing if another individual is
available (Silcox and Evans 1982). This new pair bond does not destroy
the earlier one since birds will still clump with the old partner providing
the new one is absent. Immelmann (1959) found that 35/42 pairs
reformed after periods of visual and auditory separation ranging from
two to 23 weeks, and pairs that had laid a clutch always re-paired. Simi-
larly, Clayton (1990a) found that female Australian Zebra Finches
forced to pair with Lesser Sundas males only maintained the old bond
when Australian males became available if they had already laid eggs. A
breeding attempt therefore seems to be the final confirmation of a perma-
nent pair bond.
When members of a pair are visually separated, males are more active
than females; they hop and fly more, and give more Distance Calls than
females, but both sexes show an increase in defecation rates which sug-
gests both are stressed by the separation (Butterfield 1970). Immelmann
(1959) found that when previously separated partners were allowed to
meet again in an aviary it was always the female that joined the maleo He
concluded that the pair bond was stronger in females than males. Males
also sing more Undirected Songs (Chapter 10) when separated from the
mate than in her presence. In an operant conditioning experiment involv-
ing an on-off perching response, Butterfield (1970) found that an iso-
lated male Zebra Finch, found the sight of the female partner positively
reinforcing. She was more reinforcing than other females, and most rein-
forcing when caged with another maleo Presumably, the male is anxious
to guard his female. Interestingly, females could not be similarly condi-
tioned, perhaps because they are too inactive on separation from the
partner to learn the operant response.
Auditory cues, not visual ones, are used for recognition of members of
a pair. Immelmann (1959) discovered that pairs recognised one another
irrespective of whether he painted their bills or plumage different
colours.
After a period of separation from the female partner free-living and
captive males court on reunion. This is particularly intense during the
egg-Iaying periodo Butterfield (1970) found a positive correlation
between the amount of sexual behaviour and the duration of separation,
and concluded that the pair bond has a sexual motivation. Her birds
were not in the pre-laying and laying states when tested so the courtship
on reunion cannot be regarded as a version of 'retaliatory copulation'
(see below).
In order to alternate incubation and brooding duties, Zebra Finches
must loosen the tight pair bond. This occurs gradually, and occurs in
conjunction with bonding to the nest. Initially, the partner is always in
sight, but when nest building commences the female remains at the nest
166 THE ZEBRA FINCH
si te and when the male collects nest material nearby he is often out of
sight. He now uses auditory communication to maintain continual con-
tact: Undirected Songs are sung soon after leaving the nest and Distance
Calls are given on the return trip, however, the female is silent. Should
the female leave the nest the male will follow and attempt to lead her
back inside with nest solicitation displays (below). By the start of incuba-
tion the bond has loosened to the extent that there is no need for audi-
tory contact and the partners alternate between the nest and flock and no
dumping or allopreening is observed. When the young are six to ten
days old the bond begins to be strengthened again and the pair consort
much of the time. Consorting begins earlier when high ambient tempera-
tures make brooding unnecessary over much of the day; however, when
nestlings reach ten days of age most day-time brooding has ceased and
the pair bond is complete again. Initially, auditory contact via calls is
resumed, then visual contact, and finally tactile contact where the par-
ents once again allopreen and dump. Should the nest suffer predation
the pair will begin dumping and allopreening within the hour. The first
copulations in the wild were observed on the day young fledged, but
males may initiate unsuccessful courtship with their partners even before
the end of incubation.
Nest building
Searching for a nest site
When the male leads the female in search of a suitable nest site he gives
Stack Calls (Chapter 10) much of the time. As he lands at a potential
site, these switch to Kackle Calls, and if the female shows interest he
hops to and fro from her to the site giving Ark Calls each time he lands
on the site itself. If she approaches he performs the Head-down Tail-fan
display in which he bows down and fans the tail and mandibulates at
the female while slowly pivoting the body from side to side; Ark Calls
are emitted continuously. At an old nest he may pause in the entrance
and fan his tail at the female, his white rump highlighted like a
beacon. Sometimes the male adopts the Head-down Tail-fan and nods
in the direction of the nest entrance as he mandibulates. Females are dis-
criminating about suitable sites and in most cases show no interest and
leave, whereupon the male will lead the search elsewhere. Females reject
many sites before finally settling on one. Strangely, females never lead
the nest search, a strategy that would save much time. If she is interested
in a site she may approach in a Head Tail Twist, mandibulate at the
male and hop around the site, or in and out of an old nest; this will
intensify the male's display and he will Nest Whine (Chapter 10) and
start nibbling or pushing any nest material nearby. Males often give
Undirected Songs (Chapter 10) while the female is at a potential nesting
site.
9 Social and reproductive behaviour 167
Nest ceremony
If the female accepts a site she joins the male and both sit or sprawl in
contact, side by side, on the exact location of the future nest, and inter-
mittently stretch their necks out, point their bills at one another,
mandibulate and give the Nest Whine (Figure 9.3). Later, the pair hop
to-and-fro on the nest site greeting one another with Kackle calls fol-
lowed by more Whines and Mandibulation Sprawls. Both nibble at
nearby stems or twigs, and eventually the male flies off to get the first
long grass stem for the new nest. According to Immelmann (1962a) the
installation of the first stem of a new nest is an important symbolic
moment in the nest ceremony. Amid much nest whining both male and
female position and reposition the stem and the male may even take it
away and bring it back again. More stems are fetched and the arrival
and placement are accompanied by les s and less ceremony so that the
ritualised behaviour gives over to actual nest construction. Nest cere-
monies may still given in the following days but are muted.
Nest solicitations and incipient nest ceremonies are more frequent in
Zebra Finches than in other Australian estrildines (Immelmann 1962a).
They can be heard giving Nest Whines throughout the year, even in
roosting nests. Close relatives, namely the Double-barred and the three
Poephila species, are almost identical in their performance of nest solici-
tations and ceremonies but they are more or les s restricted to the final
site chosen for the breeding nest.
Nest building-a division of labour
Nest construction follows estrildine conventions (Kunkel 1959),
although Zebra Pinches show a greater division of labour than that
found in sorne species. In Zebra Pinches both partners build the nest
Fig. 9.3 The nest ceremony is performed on the exact location of the future nest.
The partners sit or sprawl, side by side, mandibulating and whining for an extended
period before the onset of building.
168 THE ZEBRA FINCH
with the male taking the initiative. In most free-living and domesticated
pairs the male fetches the nest material and initially positions it, but the
female puts it in its final position. In a few pairs, the female may fetch
nest lining. The male will not carry stems to the nest nor do any building
if the female is not presento Nest building occurs in bouts and most mate-
rial is collected near the nest; at the Danaher colony 45% of items were
collected within 5 m, the median distance travelled was 12 m and the
maximum 87 m (Birkhead et al. 1988a). However, Immelmann (1962a)
found that most pairs in his experience collected material away from the
nest, usually flying up to 50 m away. The rate at which stems are added
varíes according to nest-building stage and individual pairs. The rate is
highest about four to five days before the first egg is laid, and sorne
Danaher pairs reached a maximum of 19 trips per hour (Birkhead et al.
1988a). At Alice Springs 1 saw one hyperactive male four days before the
first egg make 42 deliveries to the nest in one hour, in addition to per-
forming thirty-five bouts of Undirected Song, two bouts of courtship and
several brief nest ceremonies. The maximum rate of stem deliveries to the
nest observed by Immelmann (1962a) was 23 in 30 minutes. During lay--
ing the rate declines to around 3 trips per hour. Frequent performance of
Undirected Song during the searching and cOllection of nest material is
characteristic of Zebra Finches and the three species of Poephila, but
Immelmann (1962a) reports that all remaining Australian estrildines col-
lect material silently.
Although Immelmann (1962b) found that nest construction is poorly
developed in domesticated Zebra Finches, frequently, excessive nest
material is added on top of newly laid clutches making incubation
impossible.
AH nest material in the wild is collected from the ground, except feath-
ers and other lining material which may be stolen from the nests of oth-
ers, either conspecifics or heterospecifics. Previous experience with
particular materials and nesting sites can determine what will be used
(Sargent 1965). In the wild most stems are collected where they have
fallen, but sorne are snapped off with the bill at ground level or pulled
back until they break. Stems are tested by nibbling and several may be
rejected before the right one is found. Like all estrildines, stems are car-
ried by the thick end, and the flight may be laboured when holding a
long stem-difficulties can arise if they get entangled in twigs. Nest lin-
ing is bundled together and carried in the middle; feathers are preferred
and larger ones are carried by the shaft tipo
A foundation platform is built if none is presento Stems are criss-
crossed horizontally with no special fastening to the supporting twigs.
The nest is built from the bottom upwards with the material added from
the inside and worked into the surrounding twigs. The nest chamber is
built first, the floor, walls and roof are added in that order, then the nest
tu be and lining are completed together. In nest boxes or hollows the roof
9 Social and reproductive behaviour 169
and entran ce tu be are omitted. Like most estrildines, there are three basic
building movements, each with a quivering vibration that anchors the
stem: push-away, lateral-pull and pull-in (Kunkel 1959). Push-away has
several forms in the Zebra Finch; the male drags a stem into the nest, fas-
tens the end at the rear with a push up movement by stretching the legs,
neck and body and releasing the end of the stem. The head and breast
are also used to push the framework of the chamber out and up in order
to enlarge it, and the bill and head can be used to lever the lower frame-
work up. With the lateral-pull the bird, usually the female, sits in the
middle of the cup reaches forward, grasps the middle of a stem and
draws it diagonally across the body and pushing it up behind so that the
stem forms the lower walls and floor of the nest. The floor and walls of
the entrance tube are formed by the pull-in movement where the sitting
bird stretches forward, grasps the free end of a stem and draws it into
the wall.
Zebra Finches nest build actively throughout the year. If they are not
building breeding nests they are continually renovating and extending
their roosting nests, sometimes building nests they never use. Immelmann
(1962a) listed them among the three species of Australian estrildines that
have a strong building drive throughout the year.
~
Head-tail Twisls with bows
and bill wipes Hopping lo and ¡ro in
Head-tail Twists with bows
and bill wipe
/¿Stops
t
2. Song & dance
Directed Song & Hop-Pivol
Dance
~ crouches on perch
~
t
Tail-guivers
3. Mounting &
copulation
...
_---I.~ Copulation
4. Post-copulation Dlsmounts
•
Tail-quivers
head to a normal double wiping action in the air just aboye the perch;
most appear as stiff bows (Figure 9.5b). Sometimes a genuine bill wipe
after drinking appears to inadvertently initiate a courtship sequence by
stimulating the partner to begin stage 1.
Workman and Andrew (1986) noticed that courting males keep the
right flank orientated towards the female significantly more than chance
expectation. They believe this occurs because the right eye is preferred
for observing the female because of brain lateralisation processes. How-
ever, ten Cate et al. (1990) could not verify any eye bias in their tests
although both sexes had a locomotory bias to move in a clockwise direc-
tion. Further experiments are needed for a conclusive finding on eye-bias
in courtship (ten Cate 1991a).
When the female stop s waltzing and crouches on the perch the male
approaches within 20 cm and switches to his song and dance-stage 2.
He adopts the Upright Fluffed singing posture, where the head is held
172 THE ZEBRA FINCH
d e
Fig. 9.5 Courtship bchaviours and postures before, during, and after eopulation.
(a) exaggerated greeting; (b) waltz; (e) song-danee and Tail-quiver; (d) eopulation;
(e) male Tail-quiver.
erect, the bíll and taíl horizontal, and the legs bent so that the abdomen
touches the perch (Figure 9.5c). Feathers of the abdomen, flanks and ear
coverts are fluffed. As he sings he twists his head and tail towards the
female and pivots the head from si de to síde as he 'beams' the song at
9 Social and reproductive behaviour 173
her. On a straight perch he hop-pivots towards the female facing one
way then the other as he swings the body from side-to-side through 180 0
as depicted by Morris (1954). Occasionally, when they court in dense
twigs or on the ground the male hops in a semicircle around the female
only pivoting at the end of a runo At all times the tail is twisted towards
the female. Morris (1954) thought the hop-pivot was a ritualised move-
ment, but Immelmann (1959) and Kunkel (1959) considered it the nor-
mal way Zebra Finches move along a straight perch, yet no evidence was
given for either case. After a variable time the female crouches lower and
performs the Tail-quiver in which the tail is quivered up and down while
a horizontal or slightly head-down position is adopted (Figure 9.5c); this
may last ten seconds or longer until the male stops singing, mounts
and copula tes. Both male and female give the Whining Copulation
Callo Mounting and copulation only take one or two seconds. On dis-
mounting, the male often crouches and performs the Tail-quiver himself
for as long as 20 seconds (Figure 9.5e). Morris (1954) called this 'pseudo-
female' behaviour and occasionally females respond by mounting the
male, but Immelmann (1962a) thought that this was an artefact of
domestication. However, Nancy Burley (pers. comm.) saw two rever se
mountings by a pair at Alice Springs in 1986, but it has not been
recorded elsewhere.
During stage 1, both birds give a series of calls which become more
rapid and louder as courtship progresses, especially in the female. Males
mainly give Kackles interspersed with a few Distance Calls and females
mainly give Tets, but also Stacks (Chapter 10). The female calls reach a
crescendo just before the Copulatory Whines (Chapter 10). These occur
just as she begins to Tail-quiver. Song is interspersed with Distance Calls
and the male stops immediately the female solicits copulation and begins
to whine himself when his copulatory movements begin. Sonograms of
copulation show the longer, higher-pitched whine of the female followed
by the briefer, lower-pitched whine of the male on which is superim-
posed noises of rapid wing beats as he maintains balance. A bout of
Undirected Song may follow copulation.
Most courtship in the wild does not result in Tail-quivering and copu-
lation, but breaks off due to lack of female responsiveness. Courtships
initiated by the female were more likely to proceed to copulation. Some-
times females Tail-quiver before the male has sung, and he may not
mount for sorne reason, or she stop s before he is ready. Males usually
perform the song and dance, but not always, sometimes they omit the
waltz after the initial greeting. Males may chase fleeing females, adopting
the Head Tail Twists when they reach within about 40 cm; the singing
posture is adopted when they get within about 20 cm. Courtship to
non-mates may be very brief in both captivity and in the wild: simply
a song phrase or two followed by an uninvited mounting attempt,
and occasionally the song itself may be omitted. 1 have seen wild
174 THE ZEBRA FINCH
males interrupt incubation on hearing a female Distance Call, fly 50 m or
more to her, court and try to mount. In contrast, Birkhead et al. (1989)
describe a stage 2 sequence of extra-pair courtship lasting as long as two
minutes.
Occasionally, free-living Zebra Pinches perform a second type of pair
courtship whose function is unknown. During the fertile períod when
both are sitting side by side in the nest the male will suddenly hop out,
turn around in the entrance, face the sitting female and sing ro her. She
makes no overt response and the male usually rejoins her for nest cere-
momes.
The displays and sequence of female Zebra Finch courtship dosely
resemble those described by lmmelmann (1962a, 1965a) for the other
Australian estrildines which conform, more or less, to the pattern found
throughout the subfamily (Steiner 1955). Male courtship, in contrast,
mostly resembles that of the three species of Poephila, especially the
Masked Finch (Zann 1976a); stages 1 and 2 are similar in movements
and posture, but feather erection is greater in the Zebra Pinch. The
Double-barred Finch, courts differently from the Zebra Finch in that the
posture is more horizontal and the feathers more uniformly fluffed. Bill
wiping is much more frequent in Double-barred Finches, but there is little
dancing (Morris 1958). Domesticated Zebra Finches tend to abbreviate
or omit the introductory stage of courtship (Morris 1954; Immelmann
1962b), but otherwise it is the same, except for white mutants (see
below).
Pattern of pair copulation in domesticated Zebra Pinches
As part of their investigations into the relationship between copulation
and fertilisation in birds, Birkhead et al. (1989) continuously observed the
courtship and copulatory behaviour of domesticated Zebra Finches
throughout the hours of daylight by video-recording pairs (n == 10) in
single cages. Males initiated 78 % of copulations and females initiated the
remainder with Tail-quivers. Cloacal contact was more likely in the latter.
Males Tail-quivered after 62% of copulations. Female-initiated copula-
tions occurred earlier in the breeding cycle relative to egg laying than
those initiated by the maleo All copulations occurred during the female
fertile period and were most frequent in the first hour of daylight. Copu--
lations were frequent between day -5 and day 2, peaking on day -1 and
day O and dropping markedly after day 2 when the second egg was laido
Similar patterns of copulatian were found in aviary studies when birds
were observed for the first six hours of daylight over the fertile periodo
However, there were two major differences:
(1) 26% of pair mountings were disrupted by other Zebra Finches
(males disrupted mountings on 30 occasions and females on 8); and
(2) 47% oí undisturbed copulations were followed by the female
9 Social and reproductive behaviour 175
soliciting a second time, and in 30% of these another copulation
followed. Females in the cage situation, by contrast, did not invite
successive copulations.
Significance of species-specific and secondary sexual characteristics
Immelmann (1959) conducted an exhaustive series of ingenious experi-
ments on domesticated Zebra Finches in order to determine the most
important visual releasers for the elicitation of species-, sex- and age-
specific behaviours. Test birds of wild-type and white colour morphs
were given a choice of two individuals in a small cage. The stimulus indi-
viduals were either painted models, live birds of different morphs, or
white morphs painted with male or female marks or a combination of
marks. The marks that elicited species recognition were the red bill, the
eye stripe and the tail bands. These releasers stimulated initial approach
and greeting and subsequent social interactions such as dumping and
allopreening. Birds or models without head markings were ignored.
White mutants, by contrast, have only the white plumage as a species
character and use this exdusively.
Male secondary sexual marks had a strong releasing effect on female
courtship, and marks tested on their own still praduced a response. Stim-
ulated females continually landed near the most effective model and
greeted it with the Head Tail Twist, but further stage 1 courtship was
prevented by the absence of male displays-only the male courtship
dance elicited the female invitation to copulate. Presentation of painted
models and white males showed that the most effective mark was the
chestnut-coloured ear coverts, followed by the black breast band, and
the white-spotted chestnut-coloured flank marks. Darkness of the bill
was found to be of no importance (d. Burley's experiments in Chapter
11).
When the releasing effect of female appearance on male courtship was
tested, Immelmann (1959) found that the sign stimulus was a grey, cylin-
der-shaped object with a red frant. Years later, in follow up experiments,
Bischof (1980) showed that male courtship could even be elicited by two
dimensional models of females. When Immelmann (1959) offered males
a choice of females, live ones and/or models, those with pale red bills
were preferred over those with dark red bills. Males directed courtship
greetings and stage 1 courtship to the most effective models and sorne
hypersexual individuals attempted to mount them. At the Danaher
colony a continuous series of males courted (stage 2) and copulated with
freeze-dried female specimens attached to branches of bushes in the
colony (T. R. Birkhead and R. Zann, unpublished observations). Thus,
courtship in male Zebra Pinches can be elicited by two or three dimen-
sional models as long as they contain at least one correct releaser.
In eliciting courtship, Immelmann (1959) found that males were more
responsive to the appearance of the female, and while females initially
176 THE ZEBRA FINCH
chose males on the basis of their appearance, they quickly responded to
the vigour of the courtship display (Chapter 11). Similarly males could
overcome inhibitions based on incorrect bill colour if the female invited
courtship vigorously. Morris (1954), for example, believed that female
bowing was an important display in stimulating male singing and this
was later confirmed by Garson et al. (1980) using models. Bischof (1985)
used live birds to show that visual stimuli emanating from the female
were sufficient to elicit aH the male display elements in the song and
dance, but acoustical stimuli were ineffective when isolated from the
visual components of the display. The rate of male bill-wipes during the
dance was not affected by the sight and/or sound of the female, and may
simply refIect a state of heightened general arousal. In contrast, the
sounds of the female heightened the male's sexual arousal, and in doing
so, increased his rate of singing, but inexplicably, did not affect his rate
of bowing and hop-·pivoting, perhaps due to spatiallimitations in the test
situation.
White Zebra Finches are sexually monochromatic, consequently males
respond to every stranger as if it was a female (Immelmann 1959). How-
ever, the bill of males is darker than that of females, and this is the mor-
phological character used exclusively in this morph to identify the sex of
strangers. It is a relative cue, that is, males wí1l court strangers, even
other males, if no other stranger has a paler bill. Vocalisations are not
used to identify sex in face-to-face encounters in white Zebra Finches.
Moreover, the absence of secondary sexual marks causes females to flee
courting males, which in turn stimulates maJe aggression, so thar
courtship in white Zebra Finches is characterised by much fIeeing by
females and aggression by males. Eventually pairs habituate and copula-·
tion can occur, but the female courtship response remains muted.
Immelmann found that white males preferred to court white females
and models, and vice versa for wild-type subjects. However, he also
showed, through cross-fostering experiments, that this preference was a
consequence of imprinting on the rearing parents in the first month or so
of life (Cha pter 11).
Further experiments by Immelmann showed that when the pale red
bill of an adult female was painted black, the colour of fIedglings, it
inhibited male courtship for sorne time, and only strong female courtship
invitations were effective in gradually overcoming his inhibitions. Simi-
larly, an adult male given a black bill initially inhibited female courtship,
but this was soon overcome by his song and dance. Adult males first
learn to distinguish the sex of juveniles when they reach an age of 35-40
days. According to Immelmann (1959) this is not a consequence of
patches of male-specific plumage which are beginning to appear in males
at this age, but identificatíon is based on sex differences in vocalisations.
However, my analysis of sex changes in the Distance Call shows that sex
differences do not develop until 40-50 days age (Chapter 10).
9 Social and reproductive behaviour 177
In summary, Immelmann (1959) concluded that the secondary sexual
characteristics of male Zebra Finches
(1) maintain an individual distance between adult males and prevents
them fram courting one another,
(2) release the female greeting appraach flight at the onset of courtship;
and
(3) suppress the female fleeing response during his courtship dance.
While sexualIy dichromatic features are of primary importance in the
initial choice of a partner, and in stimulating the first stage of courtship,
any shortcomings in appearance can be eventually overridden by vigor-
ous sex-specific behaviours.
Extra-pair copulations
DNA fingerprinting of nestlings and their puta ti ve parents at the
Danaher colony showed that there were two cases of extra-pair paternity
among 82 birds whose maternity was confirmed (Birkhead et al. 1990).
So, for 2.4% of offspring (8% of braods and 12.5% of females) the
mother copulated with a male other than the rearing father. This inci-
dence of extra-pair paternity is lower than that found in an aviary study
of domesticated Zebra Finches (Birkhead et al. 1989) in which genetic
plumage markers indicated that extra-pair paternity occurred in 5.6%
(4/71) of offspring and 11 % of broods (2/18). Differences are prabably
due to effects of aviary confinement and domestication since the two
methods of establishing paternity have been shown to be equally reliable
(M0lIer and Birkhead 1992). Extra-pair paternity in wild birds of other
species ranges fram 0-60% (Birkhead and M011er 1992).
Theoretically, extra-pair paternity can happen in three ways-thraugh
forced and unforced extra-pair copulation, and by rapid mate switching.
All three occur in Zebra Finches and have been intensively investigated
by Birkhead and co-workers in domesticated Zebra Finches and in their
wild-caught and free-living counterparts. Rapid mate switching happens
when a female is inseminated by one male, usually the mate, but switches
to a second male with whom she pairs before the eggs are laido This can
be manipulated experimentally in domesticated Zebra Finches so that the
second male not only pairs with the female but can fertilise part of the
clutch (Birkhead et al. 1988b). In the two broods where extra-pair pater-
nity occurred at Danaher we observed the rearing parents together
before and after the eggs were laid so no mate switching occurred here,
but given the high rate of mate loss at Danaher rapid mate switching
could be a possibility. Of course, there is a reproductive cost to the
replacement male if he ends up rearing offspring of his predecessor, but
if he gets a chance to copulate with the female before the first egg is laid
he has a good chance of being the father (see below). If eggs are already
laid he should postpone pair formation.
178 THE ZEBRA FINCH
In the two cases of extra-pair paternity at Danaher we did not observe
either female making extra-pair courtships, that is, mating with non·-
mates, although we did not focus specifically on copulatory behaviour
when watching their nests. However, six weeks earlier when copulatory
behaviour of four pairs was the subject of detailed study, two of 44 copu-
lations observed were extra-pair copulations, the remainder were pair
copulations (Birkhead et al. 1988a). This proportion (4.5%) was not sig-
nificantly different fram that of offspring extra-pair paternity (2.4%) and
is consistent with the idea that extra-pair copulations are correlated with
extra-pair paternities, but this is contrary to recent findings made in a
number of other species (Dunn and Lifjeld, 1994). These two extra-pair
copulations were not forced. The male performed the song and dance
and one female invited mounting with Tail-quivers but the second did
noto Cloacal contact occurred in both cases and ejaculate was possibly
transferred.
Extra-pair courtship at the Danaher and Alice Springs colonies was
commonly observed. Over a six week period at Danaher in January and
February 1988, Birkhead et al. (1988a) observed 84 extra-pair
courtships or copulation incidents between 10 males and 15 females; 67
involved male song and dance towards the female and 15 were forced
extra-pair mountings, two of which achieved cloacal contact and pre
sumably sperm transfer. Of the forced extra-pair mountings, in 14
instances the female was retreating when the male mounted uninvited,
and in one instance the male disrupted pair copulation by knocking the
pair male off the female's back and substituted himself-all without
upsetting the female. This extraordinary method of forced extra-pair
copulation is believed to be the most successful, since females wiU other-
wise fly out fram beneath uninvited males that pounce on their backs
(Birkhead et al. 1989). Females may be more willing to participate in
extra-paír matings than they appear to be since they may be testing
males to ensure that thcy mate with thc best or most persistent individ-
ual. Males can be very persistent and will still attempt to court a female
even when they have been vigorously supplanted by her up to six or
more times. Females can prabably control the frequency and timing of
copulations despite thc constraints of mate guarding and the limited
opportunities for extra-pair matings. Studies of other monogamous
species have reached a similar conclusion (Birkhead and M011er 1993).
Why a mated female should seek, or tolerate, copulations with extra-pa¡r
males, and avoid it with others, will be considered in Chapter 1I.
Although a single copulation can fertilise an entire clutch (Birkhead et
al. 1989), sperm competition experiments in domesticated Zebra Finches
established that the timing of copulation has a profound effect on the
number of offspring fertilised. Despite the fact that female Zebra Finches
are fertile ovcr a 14-day interval the last copulation she has fertilises a
dispraportionate (50--80%) number of eggs (Birkhead et al. 1988b;
9 Social and reproductive behaviour 179
1989; Birkhead and M011er 1992). Thus, if two males inseminate a
female during a single reproductive cycle, the sperm fram the last male
have precedence over those of the first. This phenomenon is termed 'last
male sperm precedence', and an interval of about four hours duration is
needed for an effect. This accounts for the high incidence of extra-pair
paternity since a single extra-pair copulation timed to occur last in a
female's copulatory series will have a high success rateo Hence, pay-offs
can be great for a male making extra-pair copulations at the right time,
and the losses to the cuckolded male partner devastatingly high.
Males should not only carefully monitor the repraductive cycle of their
own female partner, but also those of other females breeding in the
colony at the same time. At the Danaher colony there may be up to four
or five females out of a population of 20 or 30 breeding pairs having a
fertile period at any one time, and somehow males need to detect them.
High nesting density must facilitate the monitoring of female fertile
periods (M0ller and Birkhead 1993).
In birds, the male partner has two options to minimise cuckoldry,
(1) frequent mating with the female to dilute the effect of an extra-pair
mating and to ensure that he is last; or
(2) to guard the female constantly during her fertile period to prevent
any extra-pair mating.
Zebra Finches adopt a combination of the two strategies (Birkhead et al.
1988a).
Mate guarding
At the Danaher colony males timed their extra-pair copulation attempts
for the female's fertile period and the rate peaked on day O and day 1,
the day of the first and second eggs respectively. Nevertheless, males
making breeding attempts did not seek out extra-pair copulations until
their own partners had started laying and the fertile period was more or
less complete (Birkhead et al. 1988a). The same pattern was found in
domesticated Zebra Finches (Birkhead et al. 1989). Zebra Finches of
both sexes normally guard their partner against sexual rivals at any time
of the year, and are almost always inseparable, so that in contrast to
many species there is no need for cues to initiate mate guarding. The
problem during the fertile period is that the di vis ion of labour in nest
building requires the male to leave the female at the very time when he
cannot afford to do so. The increasingly longer periods the female spends
in the nest each day after the first egg, and the increasing warmth of the
eggs, may be the cues that indica te to the male that mate guarding is no
longer necessary.
Males are constrained in their extra-pair copulation activity because
they need to guard their own partner from other males during her fertile
periodo Females unattended by their male partner were the object of
180 THE ZEBRA FINCH
86% (58/67) of extra-pair copulation attempts at Danaher so males that
fail to guard the partner adequately during the fertile period were vulner-
able to being cuckolded. A guarding male closely follows the partner
every time she lea ves the nest and attacks any males that attempt to
court her. Consequently, during the fertile period the female ínitiates sig-
nificantly more flights away from the nest site than the male, he follows a
short distance behind (Figure 9.6).
Birkhead and Fletcher (1992) investigated the possibility that a male
also refrains from cxtra-pair copulations during his partner's fertile
period because his sperm may be in short supply. He needs to mate fre--
quently with his own female to ensure his paternity of the young he helps
rear, and this has priority over extra-pair matings. It was found that a
male transferred les s sperm in an extra-paír copulation during his
partner's fertile period, than afterwards. This strongly suggests that once
a male has cea sed copulating with his own partner there are more sperm
to spare. The amount of sperm the female partner received during each
copulatíon in her fertile period was not assayed, so she too may receive
less, simply because less sperm was available, or he may husband his
supply to ensure it meets the demands of his copulating schedule. That
)0"....
- 100 _D
a'" (l)
80 ál
Ea -- 60
v, E
e (l)
0"-
"iJj (/l
o 3:
'-' o
0= IL
O ..~
~
ifI
1-4 a.
-'=
- 1-2 <f)
1-0 ::J
O
U
b.
J.:: 0-8 e. L
O
o.. k O
0-6 CL ill
::: '"
I 0.. I o..
U>
eL
OA ~
><
Ql - 0'2 ><
-'= º}
al U>
O L
:J
"}
E O O
ro '-' :2:
LL.
Doy
Fig. 9.6 (a) Mate guarding in Zebra Finches at the Danaher colony shown by the
percentage of flights away fram the nest initiated by the female (o) and by the male
(111) towards the end of the fertile periodo After day 2, the day the third egg is la id, thc
female fertile period ends, and the both birds spend more time in the nest. (b) Ratc
of extra-pair courtship performed by focal males (o) to any female, and towards
focal females (@) by any male. (From Birkhead ct al. 1988a.)
9 Social and reproductive behaviour 181
is, he may ration the number of sperm transferred to the female partner
throughout her fertile period, even though only one copulation may be
all that is necessary to fertilise the entire clutch. In order to maximise the
likelihood of his sperm fertilising his mate's eggs he must copulate fre-
quently with her and space the copulations to ensure that he has last-
male sperm precedence, and moreover, he must save enough for an
emergency retaliatory copulation if he suspects his partner of an extra-
pair copulation. In the Bengalese Finch, sperm is limited to the extent
that three copulations within three hours will deplete the supply, which
takes 24 hours to replenish, therefore the male needs to allocate sperm
carefully in order to maximise his fertilisations (Birkhead 1991). A new
method of ejaculate collection recently devised by Pellat and Birkhead
(1994) for Zebra Finches has confirmed that similar depletions occur in
the Zebra Finch (T. R. Birkhead, pers. comm.). While Birkhead et al.
(1993) found large differences (up to tenfold) in the number of sperm
stored in the glomera of individual males there was no evidence that
male Zebra Finches can adjust the size of their ejaculate in order to
ration supply of spermatozoa (Birkhead and Fletcher 1995). These males
were held under identical conditions, and presumably were in a similar
physiological condition, but this aspect was not investigated.
There are constraints on the time a male has for guarding the female
partner during her fertile period since he must leave her to collect nest
material. His schedule is hectic at this time: he brings many stems to the
nest, often performing an abbreviated Nest Ceremony with each stem
added, and on leaving the nest he usually sings sorne phrases of Undirected
Song nearby before dashing off far more stems. The male tends not to
leave the nest if the female hops outside; he will entice her back inside with
Nest Whines or lead her back into the nest. He leaves immediately she
enters, gives more Undirected Songs and is off collecting more stems. Nor-
mally, the female is fairly cooperative in respect to guarding. She waits for
the partner's return befare flying off, and often gives prolonged pre-take-
off calls as if to warn him of her impending flight. Nevertheless, in the wild
1 have seen sorne desperate males rush after their females, occasionally
while still holding onto nest material. Despite male vigilance, and apparent
female cooperation, sorne females still get off on their own during the fer-
tile period; sometimes the male fails to notice her departure and sometimes
he simply fails to follow. Males then normally leave the nest in search of
the female, giving Distance Calls continuously. Once 1 saw a male in a nest
at Alice Spring deliver a volley of loud Distance Calls into the face of his
female when she silently returned to the nest after a long absence, and on
another occasion he invited courtship and tried to copulate when she
returned. In aviaries, Birkhead et al. (1989) observed on four occasions
that males copulated forcefully with the female partner after having seen
her involved in extra-pair copulations (n = 20). Unlike normal copulations
between partners, which are unforced, the male holds the female by the
182 THE ZEBRA FINCH
young, leave the nest entran ce and follow fluttering and jumping onto
the branches of the nesting bush while giving Long Tonal Calls (Chapter
10). Should the parents fly away, the young will follow, but they are usu-
ally led back into the nest shortly after they first emerge. For the first two
or three days young are usually led back to the nest before they are fed,
but not always.
Fledging is a gradual affair; not all young fledge at once and each
spends only a short time out of the nest on its first day. Should the par··
ents re-nest, the male takes the more active role in caring for thc
fledglings, and if they venture too close to the new breeding nest they
may be attacked by the female. Nest leading disappears about four ar
five days after fledging, but they are still led back to the sleeping nest
each evening. At first the young are led in singly, but after a few days
they enter as a group.
Parent-offspring aggression in free-living Zebra Finches is rare1y
observed. Females may drive away their fledglings from the new breeding
nest, but aggression by the father towards the young was not seen. This
contrasts with the situation in caged Zebra Finches where both parents
frequently attack and drive away their young (B6hner 1983; ten Cate
1982, 1984; Clayton 1987a). This is probably an artefact of confine-
mento
Brood reduction
Even when predation is avoided, not all nestlings reach fledging age.
About nine per cent of nestlings at the Danaher colony died before band-
ing age. Ir was usualIy the smallest individuals that disappeared,
although occasionally their bodies were found beneath larger nestlings.
Whether parents purpose1y reduced the size of the brood through star··
vation of the smallest nestling is unknown. Parents were never seen to
carry nesdings, alive 01' dead, out of the nest. Sometimes live nestlings a
few days after hatching were found in the entrance tunnel or at the bot-
tom of the nesting bush, but the causes were a mystery. SimilarIy, among
domesticated Zebra Finches not aH nestlings survive. For example, Ska-·
gen (1988) found that 13% of nestlings died, most before 10 days of age,
but strangely, there was no significant difference between synchronous
and asynchronous broods. Burley (1986a) observed parents removing
struggling nestlings and dropping them outside in the aviary leaving
them to die. She believes that this was one way to manipulate the sex
ratio of offspring (Chapter 11). When Lemon (1993) experimentally
manipulated the net energy gain of breeding Zebra Finches he found that
the group with the least daily net energy reduced the brood size within
three days of hatching by actively removing the youngest. He also states
that 'broods were reduced by more subtle actions such as differential
allacation oi food between nestlings or sibling competition', but provides
no details. However, Skagen (1988) found that when food was limited
9 Social and reproductive behaviour 187
larger nestlings in a brood had faster growth rates than smaller nestlings
and concluded that parents preferred to feed larger nestlings more than
the smaller ones.
Behavioural development of young
Nestlings
Zebra Finches have two begging postures in which they gape upwards
for food. For about the first five days after hatching the nestling begs in a
typical passerine posture in which the neck is stretched upwards, the
head tilted back so that the crown touches the crop and the gap e is
directed upwards; the body, head and neck are held still, but the tongue
swings slowly fram side to side. The wings are not moved. On the fifth
day a new begging position, the head-down neck-twist, begins to replace
the head-raised position. The neck is twisted and ratated almost 180 0 to
one side so that the crown is almost level with the feet and the gape is
directed upwards (Figure 9.7). The head is swayed slowly from side to
side. The wings are not quivered as in most begging passerines, although
the wing on the opposite si de to that on which the head is twisted may
be extended out during intense begging. This begging posture is unique
to almost all estrildines. The period of transition in begging posture is
quite variable within and between broods. The rate of tongue wagging
increases with age fram around one cycle per second on day 2 up to
about four cycles at day 15, and it becomes faster during feeding bouts
(Muller and Smith 1978). Both types of begging posture are almost iden-
tical to those of the three species of Poephila (Zann 1972), but Double-
barred Finches move their heads, necks and lower mandibles more
(Immelmann 1962a, 1965a).
Fig. 9.7 A mother feeds a fledgling that has adopted the head-down neck-twist
posture.
188 THE ZEBRA FINCH
y oung nestlings will gape, and often call, when held in the hand. From
hatching until about thirteen days of age tactile and auditory stimulation
are sufficient to elidt begging in nestlings. This was shown experimen-
tally by Bischof and Lassek (1985) who found that touching the head
and bill with a stick and playing back begging sounds, stimulated gaping.
Immelmann (1962a) noticed that vibrations of the nest caused by strong
wind could also trigger long bouts of silent begging. Although young
align themselves towards the nest entrance around day 7, gaping is still
directed upwards until day 10, after which it is directed towards the par-
ents. A red-coloured pencil held at eye leve! will also stimulate begging,
but a grey-coloured one wiU not (Roper 1993). Begging in response to
foreign stimuli begins to diminish around day 13 and ceases at day 16,
beca use of an increase in fear (freezing, shuffling to the back of the nest),
as young begin to learn the features of their parents and distinguish them
from other objects (Bischof and Lassek 1985).
Nestlings do not give Begging Calls (Chapter 10) until the third day
after hatching, and only the individual being fed vocalises; from day four
to 12 other nestlings may vocalise during a feeding bout although they
are silent before and afterwards, but after day 13 they detect the
approach of the parent and begin to caH before the bout commences
(Muller and Smith 1978). Visual cues, initially provided by thc mouth
markings and tongue movements, are sufficient to e!icit parental feeding,
but acoustic stimuli soon become paramount. This was demonstrated
by playback experiments of begging sounds which not only triggered
begging behaviour in fledglings, but also stimulated the parents to
eat seeds and to provision thcm (Muller and Smith 1978). Also, thc
rate of Begging Calls emitted is affected by the level of hunger (Roper
1993).
Nestlings defecate as high up as possible around the wall of the nest
chamber so that a hard dry ring of concreted faeces forms out of normal
reach of the nestlings, which manage to stay fairly clean. By day 10 most
young back up to the entrance tunne! and defecate on the rim of the nest
cup which becomes solid with droppings. Larvae and pupae of a species
of tenebrionid beetle Platydema pascoei have been found feeding on old
faeces in nests of Zebra Finches and other species of estrildines and can
eventually clean up old nests in sorne cases (Hindwood 1951). When
inspecting nests of wild birds at night 1 have seen parents and brood
surrounded by mosquitoes with unidentified species of beetles and mires
swarming over the droppings.
By 13 days post-hatch young back up against the rear wall of the nest-
ing chamber when the nest is inspected and resist removal by hanging on
to the nest floor with their toes. After this age they force fledge, explod-
ing out of the nest with Long Tona) Calls and fluttering and flying as
best they can, followed by the parents giving Distance Calls. On the last
few days before fledging, young become active in the nest, flapping their
9 Social and reproductive behaviour 189
wings, self-preening, ruffling feathers and moving around and calling
back and forth to their parents outside.
Post-fledging behaviour
In the first few days after fledging young are inconspicuous, spending
much of the time sitting silently and motionlessly clumped together in a
row among dense bushes in the nesting colony. Parents return together at
regular intervals to feed them. At first the young reply with Long Tonal
Calls to most adult Distance Calls (Chapter 10), but after a few days
only reply to their parents and siblings. They hop to their parents while
giving Long Tonal Calls that flow over into Begging Calls then gather
either side and begin to beg in the head-down neck-twist posture. If fed
on the ground young beg in a semicircle around the parents. They thrust
themselves in front of the parents then shuffle back in the head-down
neck-twist amid shrill, irresistible begging cries. Feeding parents are most
intolerant of other young nearby that might attempt to join in and beg.
Fathers often sing Undirected Songs before and after they feed the young.
After feeding, parents seem to avoid perching near their young at this
vulnerable age, possibly because their brightly-coloured red bills might
attract predators; nevertheless, this behaviour persists in domesticated
Zebra Finches (Immelmann 1959). Several days after fledging, young fol-
low their parents to other parts of the colony, especially the social area
and local feeding areas. Here they encounter other juveniles, and soon
spend a good part of the day sharing the same bush where up to 16 ha ve
been observed at Danaher waiting for parents to return to feed them.
These juvenile groups may wander sorne distance from the colony and
join the main foraging flock if nearby.
Elements of male and female courtship behaviour first begin to occur
several weeks before the onset of juvenal moult. Subsong is evident in
young males as early as 25-30 days after hatching when they sing quietly
to themselves while perched alone in the bus hes of the colony. By day
40, when patches of sexual plumage begin to appear in most males, this
undirected type of singing is frequent, and occasionally the directed type
sung to the female during stage 2 courtship can be seen, but without the
accompanying dance, which does not occur until about day 60 or latero
Exaggerated greetings and the stage 1 waltz with its hops, bows and
bill-wipes first appear in young less than 30 days old, but again are not
regularly observed until after day 50 or 60. At first, courtship stages 1
and 2 may be directed indiscriminately to males and females, siblings
and peers alike, but by about seven weeks of age courtship is directed
exclusively to members of the opposite sexo In the wild, the first full
courtship of a female including Tail-quivering, copulation call and copu-
lation was observed in a 30-35-day-old individual when courted by an
adult male, but full courtship is generally rare until at least day 60.
Mounting and copulation by young males has not been observed befare
190 THE ZEBRA FINCH
day 90, however, it must first occur around 60 days, sínce the youngest
male detected breeding at Danaher was only 67 days old when his mate
laid her first egg. The youngest male seen carryíng a grass stem was 30
days old, and although they begin to test stems and make the three nest
building movements around this age, the earliest functional nest building
is not seen until paírs are at least 50-60 days old. The first nest cere
monies and defence of the roosting and breeding nest against conspecifics
and sparrows also occurs around 50-60 days. The time of first appear-
ance of behaviours described he re are earlier than those described for
domesticated (Immelmann 1959; Sossinka 1970) and free-living (lmmel-
mann 1962a) Zebra Finches except for subsong which was seen in cap-
ti ve males as early as 24 days after hatching. Immelmann (1962a)
believes that the components of each stagc of courtship appear first and
are later organised into functional sequences directed at members of the
opposite sexo He never saw young females courting young males with
incomplete sexual plumage, whereas two young males may court one
another and even form a pair bond, but this ruptures once most of the
sexual plumage comes out.
Development of seed-dehusking
Nelson (1993) investigated experimentally how dehusking behaviour
develops in domesticated fledglings. As expected, both maturation and
experience are important. y oung birds drop more seeds accidentally
before they can be dehusked and swallowed than do older, more mature
individuals, of equal experience. The larger the seed the more clumsy they
are. Differences in length of bill and neural coordination are believed to
be partly responsible. Nevertheless, the time it too k to learn to dehusk a
seed as fast as an adult varied from lOto 13 days, and depended on the
type of seed rather than the age of the individual. There was a tenfold
reduction in dehusking times over this interva!. In these learning tríals,
youngsters had 30 minutes dehusking practice each day, so that in the
normallaboratory situation where seed is available ad libitum, about five
to six hours of dehusking practice is the minimum needed to achieve
adult levels of competency. This would take from two to four days to
acquire in the normal captive situation and much longer in the wild
where seed is more difficult to Iocate and access and more diverse in type
and condítion. Additional experiments showed that skill acquired in
dehusking one type of seed couId be carried over to different types of
seeds, and that dehusking skills could diminish without regular practice.
Parent-offspring behaviour
Transition to independence
To investigate when young Zebra Finches forage and roost indepen-
dentIy oí their parents wc focused observations on young at the Danaher
9 Social and reproductive behaviour 191
Table 9.1 Stages leading to foraging independence in fledglings at the Danaher
colony based on data from two breeding seasons
Earliest Latest
colony for two breeding seasons, following them as far as possible from
fledging to 50 days of age and beyond (R. Zann and B. Quin, unpub-
lished observations). When banded at 12-14 days of age, all the brood
were marked with non-toxic dye to aid identification around the feeder
in the walk-in trap and at the roosting nests. Unfortunately, the early
post-fledging stage is characterised by heavy losses so sample sizes dimin-
ished dramatically after fledging (Chapter 8).
The transition from complete dependence on the parents for food to
complete independence began on the day of fledging (day 17-18) and
was completed after day 40. At first, young nibble anything: grass stems,
feathers, twigs, etc., but soon concentrate on seeds, which they pick up
and drop. By day 28, half of all fledglings observed (44 individuals from
21 broods) were foraging independently at the feeder and half were still
being fed by the parents, although not exclusively so. By day 35, only
10% were fed on sorne occasions by their parents. The timing of the
steps to feeding independence are shown in Table 9.1. Immelmann
(1962a) observed one free-living pair continue to feed their young until
they were seven or eight weeks old. With domesticated Zebra Finches,
Immelmann, (1959) found that parents normally feed their young for
two weeks after fledging; however, in exceptional circumstances it can be
extended for an additional week, but by the father only.
Within a day or two of fledging, young are led, usually by the father,
to a roosting nest. The breeding nest may be refurbished for a new
breeding attempt or it may be too fouled with droppings for further use.
The roosting nest may be in the same bush as the breeding nest or up to
100 m away if the parents relocate for another breeding attempt. The
father spends more time with the fledglings than the mother. Fledglings
are not permitted to roost in the new breeding nest unless there is an
emergency. From fledging day to day 22 they are led to the roosting nest
during the day for feeding purposes, principally by the father, and at
night for roosting up to the age of 40 days. One 40-day-old youngster
was even fed by the parents after it followed them to the nest. After 30
192 THE ZEBRA F'INCH
days of age, young begin to enter the roost independently and continue
to do so until day 48, after which they no longer roost in the family nest.
Day 42 is the age when half the young observed (n :::: 14 young from 4
broods) roost in the family nest and half roost on their own in a different
nest that may or may not be in the same bush as that of their parents.
y oung Zebra Pinches normal1y go to roost about an hour before adults,
but up to 30 days of age will enter the roosting nest during the day to
shelter from heavy rain, something adults never do. Of 16 clutches
observed, 11 roosted with both parents, three with the male only and
two with the female only. On attaining roosting independence, young
from several different clutches may share the same roosting nest.
A cautious interpretation of these observations suggests that most
Zebra Pinches reach nutritional independence from their parents by 35
days of age, but can maintain contact with the parents, as demonstrated
by their roosting association, until 48 days of age at least. We also saw
parents and young together around the feeder after this age but did not
observe any interactions or spacing that might suggest that the family
bonds still exist-their behaviour was not distinguishable from that with
other members of the flock. Neverthe1ess, it is possible that parent-off·
spring bonds are maintained in sorne subtle way beyond this age. The
bond between parents and offspring endures longer than Imme1mann
(1962a, 1965a) realised and extends well into the sensitive period for
song learning (day 35 to day 65; Chapter 10).
Nutritional independence is achieved in captive-reared Australian
Zebra Pinches by day 35, but Lesser Sundas Zebra Pinches from Timor
took 40 days (Clayton et aL 1991).
Summary
Wild Zebra Finches are highly social and flock throughout the yeaL
Aggression is common, and mostly occurs over mates and roosting sites;
no dominance orders are formed and agonistic behaviour is weakly
ritualised.
Flocks consist mostly of adult pairs and immatures. Any unpaired
adults meet and pair up in the flocks but have no special pair-formation
displays: they simply use the early stages of courtship. Pair bonds are
formed rapidly, and dumping and allopreening are the signs that a bond
is complete. Both sexes have preferences and these may form on first
meeting, but those of females normally prevail. Partners are initially cho-
sen on the basis of appearance, but subsequent behavioural interaction
9 Social and reproductive behaviour 195
determines whether a mutual bond will formo Pair bonds, which are pri-
marily sexual in motivation, are strong in breeding and non-breeding
seasons, and are constantly serviced and defended. Partners may be
stressed when forcefully separated and the male is more active than the
female in seeking reunion.
The male proposes suitable nesting sites to the female and she makes
the final decision. This is confirmed by a Nest Ceremony. The male
fetches the building material for the nest and the female works it in.
Courtship is initiated with an 'exaggerated greeting' followed by a
'waltz', after which the male sings and dances to the female who eventu-
ally invites copulation. The male dichromatic plumage elicits the female
sexual response and prevents her fleeing, but male courtship is necessary
for the female to invite copulation. Copulation is most frequent around
the day the first egg is laid and about 15 copulations occur for each
clutch of eggs. Extra-pair copulations can result in extra-pair paternity
(2.4% of offspring), and although forced copulation is possible, females
appear to control the number and timing of copulations within and out-
si de the pair bond. When two males copulate with the same female,
sperm of the last male have precedence over those of the first. Females
are fertile for a 15-day period ending before the laying of the second-Iast
egg; males guard females over this period in order to prevent copulation
with others. Enticing the female to stay in the nest is a form of mate
guarding. Females can store sperm for 10-13 days.
Zebra Finches have biparental care, but the female does the most and
loses the most weight. Under sorne conditions the size of the brood will
be reduced, but the mechanisms are largely unknown. y oung beg for
food in the typical estrildine neck -twist posture and leave the nest 17-18
days after hatching. Nutritional independence is reached around 35 days,
but roosting associations can extend another 15 days or so. The black
bill is used as a releaser to distinguish fledglings from adults. Zebra
Finches can recognise strange kin: brothers prefer each other's company
and females prefer to mate with first cousins.
10 Vocalisations
'Two species of grassfinches, the Australian Zebra Finch and the Bengalese
Finch, possess an instinctive song model that consists of only a few elementso
Song develops in the young through listening to the species-specific song:
however, if reared by a heterospecific it will give the foster-father's song and
completely mas k the instinctive component.'
Translated from K Immelmanl1 19670
The Australian subspecies of Zebra Finch is highly vocal, and in the wild
one usually hears them before seeing themo They have ten or eleven dis-
tinct ca lIs in addition to thc songo The Lesser Sundas subspecies may
ha ve a similar number, but to date, only the song and three caUs have
been describedo In general, aviary-bred Lesser Sundas Zebra Finches are
less vocal (both songs and calls) than their Australian counterparts in
both aviaries and cages. The sounds of each subspecies are qualitatively
distinct: Lesser Sundas birds have a high-pitched whistle while Australian
birds have a heavy nasal twango
Immelmann (1 962a, 1 965a) was the first to describe the vocalisations
of Zebra Finches, and those of the other Australian estrildines, which he
did without the benefit of tape-recorders and sound spectrograms. When
analysed on a sound spectrograph, the vocalisations of Zebra Finches,
like most estrildines, display a complex structure in which the sound
energy is concentrated into multiple bands that occur at harmonic inter
vals, and are referred to as 'harmonics' by most workers. The lowest har
monic is the fundamental. The Australian subspecies has more
harmonics stacked into its sounds than does the Lesser Sundas subspecies
and this accounts for the nasal-sounding timbre. The Double-barred
Finch and the Masked Finch also ha ve many harmonics in their sounds
and in the field their calls can be mistaken momentarily for those of
Zebra Pinches.
Cal! repertoire
When active, Australian Zebra Finches caH much of the time, except
when feeding. The three most frequent calls are the 'Distance' Call, rhe
'Tet' and the 'Stack' (Figure 10.1), which are given on the move any·
where at anytime. Around nesting colonies less-frequently uttered sounds
are also heard. During the breeding season members of a paír give spe-
cial calls at the outset of nest-buildíng ('Kackles', 'Arks' and 'Whines');
ca lIs are given beforc ('Wssts') and during combat ('Distress' Cries)o
10 Vocalisations 197
Young give a range of Begging Calls and have their own distance call
(Long Tonal Call). Parents have a call (the 'Thuk') that warns of the
approach of potential predators of nests and young. The 'Undirected
Song' is given at anytime of the year in both breeding and non-breeding
seasons while 'Directed Song' is only given during pre-coital courtship
and during pair formation.
The contexts, and putative messages and functions of the various
vocalisations emitted by the Australian Zebra Finch are summarised in
Table 10.1.
Tet (Figure 10. la)
A series of soft 'tet-tet' sounds are emitted during most hopping
movements, either between perches or when on the ground. These are
probably the sounds most frequently uttered by Zebra Finches, and in
many situations form a soft background hum in which other calls are
embedded. That is, they are not directed at specific individuals and do
not stimulate specific replies. Tets do not carry beyond about 5 m and
are mostly heard by the mate or family members nearby although in
captivity the calls are given automatically by visually and acoustically
isolated individuals. Tets are given frequently when birds are excited and
the series beco mes most rapid just before flight. In this way, take-off may
be coordinated among mates, family and flock members. The Double-
barred Finch and the three species of Poephila also use pre-take-off
warning calls (Immelmann 1962a; Zann 1975).
Tets are stereotyped in physical structure. Sonograms show apile of
crowded chevron-shaped harmonics with a duration ranging from 46-52
ms with a mean of 50 ms; the fundamental frequency ranges from 0.40
to 0.48 kHz. There are no sex differences, but birds from central
Australia have significantly longer calls of lower fundamental frequency
than birds from northern Victoria (Zann 1993a). Tets from Lesser
Sundas birds are longer and ha ve a higher fundamental frequency than
those of Australian birds (Figure 10.la vs. Figure 10.2a).
Distance Call (Figure 10. lb)
The short, ringing 'tia' of the males and the longer, more nasal and flat-
ter, 'tiaah' of the females are the loudest and most penetrating sounds
given by Australian Zebra Finches. The bill is barely opened when softer
versions of the call are emitted, but opened more for louder ones. The
Distance Call in the Lesser Sundas birds is a loud high-pitched whistle in
both sexes. In the field the Distance Call of the Australian subspecies
may be heard from 80-100 m away. It is the most characteristic sound
heard from wild Zebra Finches and is most persistent in scattered and
isolated birds. It is emitted in a variety of contexts: on take-off, in flight,
during mild alarm, greeting newcomers, stage 1 of courtship, between
bouts of singing, and between bouts of feeding of the young. Partners
Table 10.1 Commonly uttered vocalisations of Australian Zebra Finches and their putative messages and functions.
Tet Call 10.la (1) Most hopping movements 'Partnerlfami!y: Close contact.
(Communication Call) (2) Pre-take-off keep close!'
(3) Between songs and song phrases ?
Distance Cal! 10.lb (1) Isolated from partner and/or flock 'Jo in me!' (1) Distant contact and localisation
(Long Ca!! Identity Call) (2) Take-off '1 am here!' (2) Identity
(3) Stage 1 courtship 'It's me!' (3) Alarm
(4) Sudden dangerlexcitement 'Flee!'
(5) Greeting
Stack Call lO.le About to take-off and to land. 'Wateh/join me?' Intention
Wsst Call 10.3a About to supplant. (Movel' Intimidation
Thuk Cal! 10.3a Approach of potential danger from birds 'Prepare lO flee!' Warning to young
reluctant to flee
Distress Cry 10.3a (1) lntense com ba t Painlfear
(2) Contact with potential predator.
Kackle Call 1O.3b Landing next to partner near 'Greeting: follow me '\!est-site attraction?
potential nesting site to a nesting site!'
Ark Call 1O.3b Landing at potential nesting site. 'J oin me here a t this Nest-site solicitatíon
nesting site!'
Whine Call 10.3b (1) Nest ceremony: partners at nesting 'Nest here!' (1) Nest-site bonding.
site or in a nest
(2) In nest matel young outside 'Come to me in the nest!' (2) Nest-attraction.
Copularion Call During copulation 'Do not flee!' Sexual appeasement
Begging Calls 10.3c Begging for Eood by young 'Feed me!' Food solicitation.
Long Tonal Call 10.3c Isolated from parents and/or sibs 'Join me!' Localisation.
Direcred Song Direeted at female during pre-eoita! courtship 'Let's copulate!' Sexual
Undirecred Song 10.6 (1) Near nest, female inside 'Stay in the nest!' (1) Nest bondinglguarding
(2) In/near f10ek/nesting eolony 'I could be available' (2) Sexual ad vertising?
(3) Unmated males in f10ek 'I'm unmated!' (3) Weak sexual?
(4) Visual isolation from conspecifics ?
10 Vocalisations 199
(a) 'Tets'
males females
-q:
'.o.»._:.. ..............:..L .. .
MtW
4
i_
,-
,'"w:&;
0.5 s
+~~tmt::
o :k t~~;;::
Fig. 10.1 Sonograms (wide-band filter) of the three most frequently uttered vocali-
sations emitted by Australian Zebra Finches. Recordings of wild birds from northern
Victoria and Alice Springs are shown. Each sonogram is from a different individual.
200 THE ZEBRA FINCH
o
(e) Distance Calls females
kHz
8
0.5 s
4
males
o
Fig. 10.2 Sonograms of three of the most frequently uttered calls made by the
Lesser Sundas Zebra Finches. Recordings were made from captive-bred Timor birds,
and each call shown, except for 'Thuks' is from a different individual.
10 Vocalisations 2 O1
may exchange Distance Calls with one another sorne distance away
when one is in the nest and the other is flying to, or from, the nest. Males
call more frequently than females (Zann 1984). In captivity the Distance
Call is elicited by visually isolating a bird from its mate or companions.
It has a high reply rate especially between mates and members of the
immediate fami1y and flock, but newcomers are also greeted from afar
with a loud exchange of Distance Calls. When flocks of several hundred
or more assemble, prolonged volleys of Distance Calls can be irritating
to people. The call appears to function as an identity call, a lost call and
a flight call and is the most important one given by the species. Details of
sexual and geographic variation in structure, ontogeny and control of the
Distance Call are given latero
Stack (Figure 10. le)
Louder, longer and higher pitched than Tets, but softer, shorter and
lower pitched than Distance Calls, Stacks are emitted at the moment of
take-off. Thus, they are given after the series of pre-take-off Tets and
usually precede the Distance Call which is given a second or so after
take-off. Stacks are also given in flight, especially when hovering, or
when birds are hesitant about flying down alone to feed or bathe, or to a
nest at incubation change-over. Males leading females from one bush to
the next in search of a favourable site for nesting give Stacks. There is
little frequency modulation so that their harmonics are flat, horizontal
bands, although there may be sorne modulation at the onset. There are
no obvious sex differences in structure. Birds isolated visually and audi-
torily in a cage give this call more often than Tets and Distance Calls and
hundreds can be emitted in an hour. The call is structurally stereotyped
in males, but female calls are variable in duration, and this may be a
reflection of their leve1 of motivation. Female Stacks have a bimodal
distribution in duration with a set of brief calls 0.05-0.1 s, and another
set around 0.15 s duration; both are identical in harmonic structure to
Distance Calls (0.2-0.3 s) but with a slight1y lower pitch.
Wsst (Figure 10.3a)
A hissing 'wsst' sound 'like the sud den ripping of a piece of cotton c1oth'
(Imme1mann 1965a) is given by Zebra Finches a fraction of a second
before they supplant (Chapter 9) an enemy from a percho The calls are
given singly and the sonogram shows a burst of noise of low frequency.
This is the noisiest call given by Zebra Finches and the last sounds heard
at night as birds go to roost.
Thuks (Figure 10.3a)
A short, thick-sounding 'thuk' is punched out by parents at the approach
of potential predators of nests or newly fledged young. It is an alert to
get ready to flee when danger is not immediate but suspicion is aroused.
202 THE ZEBRA FINCH
Fig. 10.3 Sonograms of nine less-eommonly uttered voealisations from wild Aus-
tralian Zebra Finehes. Numbers below eaeh sonogram in (e) give the age in days
after hatehing when the reeordings were made.
10 Vocalisations 203
The threshold for caUing is lowest in parents on the day their young
fledge. Immelmann (1962a) observed that the frequency of calling begins
to increase when young are about seven days old. By this age the Begging
Calls are lo ud and penetrating and can attract the attention of predators.
Typically, the parent waiting outside the nest gives the call as its partner
feeds the young inside. The frequency of calling returns to normal levels
about 10 days after fledging.
Kackles, Arks and Whines (Figure 10.3b)
These three calls are given at a potential nesting site by pairs searching
for a suitable place to breed. Males, who take the initiative in nest site
searching, normally call the most, but females also make identical calls.
As the pair fly from bush to bush checking out old nests, the male gives
Stack calls as the female foUows. When he finds a site he attempts to lead
the female to it; first, he lands next to her and gives a series of loud,
raspy Kackle calls, then he hops to the si te still giving Kackles, and once
he lands on it he switches to Arks, which are low, long harsh, 'ark-ark'
sounds. These calls are directed at the female and as he caUs he will fan
the tail slightly, bow down and mandibulate (nibble) towards her. The
intensity of the display increases if the female looks at him or moves
towards him. If she comes close he may switch over to the Whine. This is
a long drawn out 'pleading' sound, like a small child whining for some-
thing and should the female find the site acceptable both partners wiU
whine together as they sprawl si de by side in contact and mandibulate at
each other (Chapter 9). The Whine is the longest call given by Zebra
Pinches. The three calls may intergrade as the male moves from the part-
ner to the nest site. The calling birds continually monitors its effect on
the partner and modifies them accordingly. Kackles and Arks are also
given by the male as he brings stems to the nest during building and may
also be heard in the evening when birds are going to roost.
Copulation Call
During copulation both male and female emit a short series of whine-like
caUs. The female call begins as soon as she solicits mounting with her
Tail-quivering display (Chapter 9) and the male begins when his copula-
tory movements commence. Clear sound recordings are difficult to make
with both caUing simultaneously and noisy wing flapping by the male,
consequently no sonograms have been presented. My imprinted Zebra
Finch, Fred (see Preface), gave short whines when he copulated on my
finger.
Distress Cries (Figures 10.3a,c)
When distressed or in pain Zebra Pinches may utter a high-pitched
shriek with a wide-opened bill. It is most commonly heard when young
birds are handled close to fledging. Adults rarely emit the call in these
204 THE ZEBRA FINCH
circumstances, and restríct the call to when in pain during vicious fights.
Parents respond to the distress críes of their young with Distance Calls
and by fIying to them. Opponents, however, seem to ignore the callo
Distance Cal!
This is the only long-distance signal given by Zebra Finches and
other species of estrildines since almost alI members of the sub-
family lack a loud territorial songo The Zebra Finch Distance CalI has
a number of long-distance signalIing functions. Moreover, experi-
mental studies have shown it to be unique in its ontogeny and physio-
logical control.
Structural Variation
When analysed on a sonograph physical variation is found in the struc-
ture of Distance CalIs given by Australian Zebra Finches between the
sexes and among individuals, colonies, and geographic localities; further-
more, the calls are clearly distinguishable from those of the Lesser Sun-
das subspecies (Zann 1984).
Individual variation
Each individual sampled from the same colony reproduces its Distance
Call with great consistency such that a nalve independent observer can
quickly classify sonograms by eye according to features unique to each
individual with 100% success. Univariate and multivariate statistical
analysis of various morphological characters of sonograms of Distance
206 THE ZEBRA FINCH
Ca lIs can also distinguish individuals. Thus, the Distance Call is struc--
turaHy unique to each individual. Variation in the Distance Call is
greater among males than females (Zann 1984).
Sexual variation
Sonograms of male calls of the Australian subspecies are immediately dis--
tinguishable from those of females beca use they are more complexo While
the female call is a long complex tone in which the multiple harmonics
show little frequency modulation and so appear as flattened or slíghtly
arching bands on the sonograms, the male Distance Call has harmonics
with two sharply contrasting, temporally distinct features. The first part is
a complex tone in which the harmonics have a slight upward inflection in
frequency, followed by a second part where the harmonics sweep dramat-
ically downwards (Figure 10.lb). The first part of the caH is termed the
tonal component (T), beca use of its structure, and the second, the noise
component (N), because it gives the caH a harsh, grating qua lit y (Zann
1984). Simpson and Vicario (1990) have termed the second component
the 'rapid frequency modulated segment'. Ninety-nine per cent of free-liy-
ing males (n = 258) have Distance Ca lIs of this type (TN) wíth the noise
component ranging fram 50% to 90% of the total duratíon of the call;
the exceptional individuals may omit the tonal component (all N), ar
have the noise component first (NT), ar sorne have two noise components
(NTN), but they never omit the noise component (i.e. aH T). Grass sex
differences in calI structure are not evident in the Distance Ca lIs of the
Lesser Sundas Zebra Finches beca use the males do not have a two-com-
ponent callo The whistle-like caH has harmonics with an upslur-downslur
configuration which is more ar less identical in both sexes.
Additional sex differences in caH structure become evident when the
sonograms are measured (Zann 1984). Male caHs are significantly briefer
than female ca lis in both subspecies: Lesser Sundas subspecies fram
Timar: 0.13 ± 0.03 s eX ± s.d.) (males) vs. 0.17 ± 0.01 s (females); Aus-
tralian subspecies fram northern Victoria: 0.14 ± 0.02 s vs. 0.20 ± 0.03 s.
The frequency of the fundamental, the lowest harmonic band, also diE-
fered between the sexes in the Australian subspecies (1.15 ± 0.23 kHz
(males) vs. 0.62 ± 0.06 kHz (females)), but not in the Lesser Sundas sub-
species (Lesser Sundas subspecies: 2.59 ± 0.65 kHz (males) vs. 2.08 ±
0.42 kHz (females)). Thus, more harmonics can be stacked into Aus-
tralian female ca lis than male caHs and this accounts for the more nasal
timbre. The frequency of the harmonic with the greatest amount of
sound energy, the emphasised harmonic, did not differ significantly
between the sexes in either subspecies (Lesser Sundas subspecies: 4.47 ±
0.75 kHz; Australian subspecies: 3.6 ± 0.44 kHz).
Sexual differentiation inthe Distance CalIs of estrildines equivalent to
that found in the Australian Zebra Finch have, to date, only been
described for the Bengalese Finch (Zann 1985b; Yoneda and Okanoya
10 Vocalisations 207
1991). Sexual differentiation in Distance Calls of Double-barred Finches
have not be en described, and only slight quantitative differences in
frequency and duration exist in the three Poephila species (Zann 1975).
Güttinger and Nicolai (1973) found no sex differences in the Distance
Calls of 76 species of African and Asian estrildines.
Geographic variation
Female Distance Calls from individuals belonging to 10 colonies over
southeastern Australia differed significantly in duration of the call, but
they did not differ in the frequency of the fundamental, nor in the
frequency of the most-strongly emphasised harmonic. Males differed in
duration of the call and the duration of the tonal component, the funda-
mental frequency and the emphasised frequency (Zann 1984). Gross geo-
graphic differences in the sonograms of Distance Calls are evident, but
more so in males than females.
Surprisingly, Clayton (Clayton et al. 1991) could find no geographic
variation in the Distance Calls of the Lesser Sundas subspecies. The
absence of call differentiation may indicate that populations are not
sufficiently isolated or possibly that calls in this subspecies have rigid
developmental programs with little gene tic variation.
Subspecific variation (Figure 10.lb and 10.2c,d)
Inspection of sonograms of the two subspecies reveals gross differences
in call morphology. There are more harmonics in the Australian sub-
species and only males of this subspecies have the diagnostic noise com-
ponent in the second part of the callo Measurements show that males of
the two subspecies do not differ in total duration of the sound, but the
fundamental and the emphasised harmonic are higher in the Lesser
Sundas males. Calls of Australian females are longer, but the fundamen-
tal frequency and that of the emphasised harmonic are significantly
lower (Zann 1984; Clayton et al. 1991).
Development
The adult Distance Call of Australian Zebra Finches is not fully devel-
oped in males until 60-80 days after hatching, while the female call
matures much earlier (Zann 1985b). Calls of both sexes originate from
the Long Tonal Call of fledglings. Between day 22 and day 30 the Long
Tonal call becomes briefer, and the harmonics more modulated and
amplified, so that by day 35-40 the adult version of the female Distance
Call is complete. Around day 40-50 the flat harmonics of the male call
begin to display rapid changes in frequency modulation where there is a
steep upslur at the onset, a steep downslur at the offset, and a sustained
unmodulated section of variable duration in between. By day 60, and
frequently much earlier, amplitude modulation produces the essential
elements of the adult Distance Call: the upslur at the onset disappears, so
208 THE ZEBRA FINCH
that the middle unmodulated section begins the call and forms the tonal
component and the rapid downward sweep of the offset forms the noise
component. Next, the amplitude of the higher frequencies of the caH 1S
greatly reduced so that the downward sweeping harmonics descend from
non-existent high tonal components as if cut off by a filter. Concomitant
with the changes in the configuration of the harmonics there is a reduc-
tion in number from six or seven harmonics at day 35 to three or four at
day 60 with an associated increase in frequency of the fundamental from
0.6 kHz to around 1.2 kHz. By day 60-70 the male Distance Call has
developed two components and has become higher-pitched than that of
the female.
Cross-fostering
When foster-reared by Bengalese Finches from hatching to day 40 or day
60 the Distance Calls of female Australian Zebra Finches at 100 days of
age were identical to those of their normal1y reared sisters except that the
frequency of the fundamental was slightly, but significantly, higher
(Zann 1985b). In contrast, every foster-reared male completely lacked
the noise component at the end of the call, and in 84 % of cases the tonal
call superficially resembled the pure tonal calls made by females, but
sounded less nasal (e.g. Figure 10.4 foster-reared son number 2). In other
foster-reared males, the Distance Call was identical to that of their
foster-mother (e.g. Figure 10.4 foster-reared son number 1) or foster--
father, or resembled the harmonic configuration of their foster-parents'
calls to sorne degree. The pure tonal caH of foster-reared males was nor
mal in all respects but for the absence of the noise component.
The length of foster-rearing (40 vs. 60 days) had no significant effect
on the development of the Distance Calls in the experimental and control
graups. Thus young males can learn the features of the noise component
of the Distance Call before 40 days of age, and evidence suggests that the
father serves as the model for this information (Zann 1985b). Foster-
reared males exposed to the normal male Distance Cal! after day 40 did
not develop a noise component into their cal!. This suggests that after the
first 40 days of life the sensitive phase for learning the noise component
of the Distance Call in male Zebra Finches ends. However, Slater and
Jones (1995) found it was possible for a Zebra Finch kept with his father
until 35 days of age to learn the Distance Call from a second male if sub-
sequently confined with him until 70 days of age (5 males learned from
the father and 12 males learned fram the tutor). Therefore, it is likely
that the sensitive phase for caH learning extends beyond 40 days of age
in sorne circumstances.
When Price (1979) conducted sound-isolation and deafening experi-
ments on nestling Zebra Finches, he found no differences in structure
of Tet, Stack and Distance Calls between treatments and controls and
concluded that auditory experience was unnecessary for normal develop-
10 Vocalisations 209
wild-caught normally-reared foster-reared foster-reared
kHz
father son son # 1 son # 2
8
0.5 s
4
Fig. lOA Effects of foster-rearing on the structure of the adult Distance Call in
Australian Zebra Finches. Distance Calls of a family of adult Zebra Finches are
shown. Both mother and father were wild-caught birds and the four offspring were
reared in captivity. The three foster-reared offspring were reared from the egg stage
by Bengalese Finches for the first 40 after hatching and the normally-reared son
stayed with his natural Zebra Finch parents until 40 days. Foster-reared son # 1
imitated the Distance Call of his foster-mother, while foster-reared son # 2 did not
learn from his foster-father or mother but gave an aberrant Distance Call more like
Zebra Finches than Bengalese Finches. The foster-reared daughter's call strongly
resembles her natural Zebra Finch mother, not her foster-mother.
210 THE ZEBRA FINCH
ment of these calIs. Unfortunately, all his contrals had the typical domes-
ticated version of the Distance Call, which has no noise component, so
differences between the treatments were impossible to detecto
In summary, these cross-fostering experiments show that one part of
the developmental program of the. male Distance Call is o!Jen to environ-
mental influences, but litde of the female pragram is open. The male pra-
gram specifies the total duratíon of the call, the number of harmonics,
the frequency of the fundamental and the emphasised harmonic. How-
ever, specifications of the harmonic configuration need information fram
the external model, namely, the father; he is needed to pravide program
details on where to modulate the harmonics and how much amplitude to
apply. In sorne males the specification 'learn the father's Distance Call'
appears to override everything, even the inherited specifications, so that
when crass··fostered to another species sorne males will learn the whole
Distance Call of the heterospecific. At least 30% of normally reared
males gave identical Distance Calls to their fathers. In one case, two
brothers were foster-reared in the same nest, one copied the foster··
father's Distance Call and the other gave the deprived Zebra Finch ver-
sion. Ir is not known whether these two developmental pragrams are
alternative strategies in the acquisition of the male Distance Call 01' if
they are just different responses to post-fledging environments, for exam-
pIe, the quality of parental care, the frequency of parental calling and the
number and sex of siblings.
Effects of domestication
Whereas only 1 % of wild Zebra Finches have aberrant Distance Calls,
the incidence increases to a bout 3 % (2/61) in offspring of wild-caught
birds. These birds may ha ve either no noise component (a11 T) 01' no
tonal component (al! N) (Zann 1984),
Effects of domestication on the structure of male Distance Calls has
been investigated in Australian stocks by Carr (1982) and in Japanese
and American stocks by Okanoya et al. (1993). In domesticated Aus-
tralian stocks, the incidence of thc components occurring in the same
order as that found in wild birds (TN) varied fram 17% to 23%, and in
the remaining birds the N component was always presento In contrast,
20% of the American and Japanese stocks had no N component at aH;
50% of American males had the components in reverse order, NT, and
50% of the Japanese had the species-typical TN.
Carr (1982) also found that sons of domesticated birds copied their
fathers' Distance Calls precisely, a finding confirmed by Simpson and
Vicario (1990). Moreover, the effects of cross-fostering of domesticated
Zebra Finches are identical to those found with offspring of wild-caught
birds.
The aviary environment, rather than regimes of artificial selection, is
responsible for the production of the aberrations in the domesticated
10 Vocalisations 211
male Zebra Finch Distance Callo When male offspring of wild-caught
and domesticated Zebra Pinches are cross-fostered to one another,
genetic sons of wild-caughts learn the aberrant Distance Call of their
domesticated foster-father and his genetic sons learn the normal Distance
Call from their wild-caught foster-father (Figure 10.5). Precisely what
features of the captive environment are responsible for the increasing
error rate in learning the noise component are unknown, nor is it known
what selective forces in the wild make the structure of the Distance Call
so stereotyped.
Conclusion
In most species of passerines the development of calls, as distinct from
songs, has a closed developmental program, in the sen se that specific
social experiences are unnecessary for the development of species-typical
calls. The Distance Call of male Australian Zebra Finches is one of the
exceptions to this generalisation, since foster-rearing of a number of
Wild-caught Domesticated
father father
%®1W%
~
""
~ ==:::............
........................................... ;;;
,,"""
" ........................-
...................
,r-*' ., . . . . . . .........._...
."".
.,;........
; ~,
~
"............ ...~ ------_-...
~
NormaI-reared Foster-reared Foster-reared Foster-reared
F-l son #1 domesticated son F-l son #2 F-1 son #3
other species of estrildines has not shown any evidence of caH learning
(Güttinger and Nicolai 1973). Recently, however Yoneda and Okanoya
(1991) found that male Bengalese Finches must also learn their Distance
Call in a similar way to that of Zebra Finches.
Neural control
Neural control of Distance Calls differs between the sexes in the Aus-
tralian Zebra Finch (Simpson and Vicario 1990). The simpler female
Distance Call has a less complex, more basic, neural pathway, which is
possibly located in the brain stem where it is associated with the centre
that controls respiratory patterning. The more complex mal e Distance
Call, in contrast, is controlled by the same complex neural pathway in
the telencephalon that controls song production (see below). Simpson
and Vicario (1990) established the function of this pathway in the
production of the maJe Distance Call by destroying or cutting the
different components. The noise segment, the short duration and the
high fundamental frequency were al! lost when the vital neura) centres
were lesioned, resulting in a Distance Call almost identical in physical
structure to that of the female. Destruction of the pathway in females
had no effect 00 the structure of the female Distance Call, proving that
this pathway had no role in its control. Simpson and Vicario (1990)
postulate that the maJe has two co-existing neural pathways for the pro
duction of the Distance Call: the song control circuit, which produces the
learned Distance Call with its male·-specific duration and fundamental
frequency, and the more basic brain stem pathway that produces the
female callo y oung males produce the female version of the Distance Call
before the song circuit is fully functional at around 60 days of age, and
older males faH back on the female version of the Distance Call and
its basic neural control if experimentally induced lesions damage the
functions of the song circuito If oestrogen is implanted in female nestlings
soon after hatching their brains wiU also develop the song control circuir
and they are then capable of producing the learned male-like Distance
Calls (Símpson and Vicario 1991a,b).
Individual recognition
Theoretically, the acoustic structure of the Distance Calls of Australian
Zebra Finches could impart information on the caller's age, sex and geo-
graphical origins as well as its individual identity. Two unpublished
experimental studies suggest that the Distance Call enables recognition
of mates, parents and offspring. In domesticated Zebra Finches, Silcox
(1979) found that females discriminated between the Distance Calls of
their partners and those of other males whereas males responded with
equal fervour to aH females. Using second and third generatíon offspring
of wild-caught Zebra Finches, MeIntosh (1983) found that members of a
pair recognised thc Distance Calls of their partners in simultaneous play·
10 Vocalisations 213
back experiments where calls of birds of the same sex with whom they
were familiar served as controls. Under the artificial conditions of this
experiment, where the dynamic exchange of Distance Calls between
stimulus and test subjects was not possible, females showed a stronger
preference towards their mate's Distance Call than did the males. How-
ever, males called more than females during playback, and were more
active, but females altered their rate of calling more than that of males.
Butterfield (1970) made the same observation when members of a pair
were visually, but not acoustically isolated from one another. McIntosh
hypothesised that members of a pair relocate one another in the follow-
ing way: the male calls more than the female when separated and moves
around actively searching for her; when he comes within earshot she
recognises his call and responds with her own Distance Call but remains
stationary; the male homes in on the female that alters her call rate in
response to his Distance Calls. This explanation is consistent with the
finding that free-living males call more than females and that the physical
structure of the male call is more individually distinct and thus easier to
recognise than that of the female.
In other playback experiments McIntosh showed that parents discrimi-
nated the Long Tonal/Distance Calls of their own fledglings from those
of others. Mother and father responded equally and responded to the
calls of birds as young as 20 days of age, but the response waned when
the young reached 32-35 days of age, and ceased completely soon
afterwards.
Fledglings also recognise the Distance Calls of their parents: they sud-
denly cease moving and reduce their rate of Tet Calling when their par-
ents call but they ignore the Distance Calls of others. McIntosh believes
that the parental Distance Calls have a tranquillising effect on the newly
fledged young under these experimental conditions. Naturalistic observa-
tions by Immelmann (1962a) suggest the opposite response; young call in
reply to the Distance Call of all adults for the first five to seven days after
fledging, but soon learn to recognise calls of both parents and restrict
their replies to them. According to Immelmann, young can pick out their
parents from a multitude of callers at distances of up to 100 m, or so,
away.
Song
Of all Zebra Finch behaviours we know most about song, namely that
series of various vocal utterances normally sung by the male to the
female during precopulatory courtship. It has been the subject of intense
investigation since Immelmann's (1965c, 1967) pioneering studies into
its development. Domesticated Australian Zebra Finches have become
the principal focus of investigations into development, control, percep-
tion and function of song, and, recently, variation and structure of song
214 THE ZEBRA FlNCH
in wild birds has also been investigated. Unfortunately, no studies of
singing have been made on wild Lesser Sundas Zebra Finches, although
sorne laboratory work has been conducted in recent years.
Zebra Finches sing a soft song that sounds squeaky and cheerful to the
human ear and, perhaps, more mechanical than musical. Parts have a
strange ventriloquial quality that makes singers hard for humans to
localise. Song of the Lesser Sundas subspecies is softer and noisier than
that of the Australian subspecies. Singing can be heard in wild Australian
Zebra Finches throughout the year, irrespective of whether they are
breeding or noto However, wild Lesser Sundas Zebra Finches sing less
often; for example, Clayton et al. (1991) rarely heard them sing during
their field study that spanned part of the breeding periodo In Australia,
sorne parts of the song can be heard more than 20 m away, depending
on the weather, but one needs to be much doser to hear the complete
songo Distance Calls, both lo ud and soft, are embedded among the notes
that make up the song, but these cannot be distinguished in the songs of
the Lesser Sundas subspecies. Careful listening will also allow one to dis
tinguish songs of difierent individuals since each is unique, an observa"
tion first made by Morris (1954).
The sound and organisation of Zebra Finch song is similar to that of
the Double-barred Finch (Hall 1962) and the Masked Finch (Zann
1976b). Both species have harmonically rich elements and there is no
interruption between the phrases. In contrast the Long-tailed and Black~
throated Finches have a number of tonal elements in their songs and
pauses between phrases.
Structure
In both subspecies of Zebra Finches notes occur in sequences that are
repeated throughout each song performance and constitute phrases, or
verses. Sonograms of two wild Australian Zebra Finches are shown in
Figure 10.6 and three captive-bred Zebra Finches from Timor shown in
Figure 10.7.
Unfortunately, investigators of Zebra Finch song have not standard··
ised their terminology. The smallest components, or units, are called
elements (syllables, notes). On sonograms elements appear as morpho-
10gical1y discrete black tracings that are separated from their neighbours
by intervals without tracings, namely gaps of silence that range from 5 to
10 ms in duration. In sorne instances elements are not separated by gaps
of silence but by sudden changes in amplitude (Williams and McKibben
1992). Elements are not merely descriptive units but functional ones as
well, since Cynx (1990) has shown that singing is never interrupted in
the middle of elements, but only in the silent intervals. Consequently ele-
ments as depicted by tracings on sonograms are the vocal units of pro-
duction in Zebra Finch songo However, when elements are copied the
element itself and the silent interval preceding it are copied as one unit
10 Vocalisations 215
(a)
o
11 12 N-N H N-5-T 51 OC 52 53 11 12 N-N H N-5-T 51 DC 52 53
(b)
introduction phrase 1 phra8e 2
kHz --------------,------------------------,------------------------,
8 \ I !
0.58
4
o
- - -- -- - -- -- --
11 L H N-5 DC 12 L H N-5 DC
Fig. 10.6 Sonograms of songs fram two wild Zebra Finches fram different parts of
the Australian distribution: (a) Alice Springs, and (b) northern Victoria. Both songs
are of the undirected type in which a female is not involved, and show two succes-
sive song phrases. The start of singing is shown in male (b) where a couple of intra-
ductory notes precede the first phrase. The letter code beneath each note or
element indicates its type, and the number refers to the version within that type: 1,
Intraductory; L, Ladder; H, High; N-S, Noise-structure; DC, Distance Call; N-N,
Noise-noise; N-S-T, Noise-structure Tone; S, Stack.
(Williams and Staples 1992), hence, strictly speaking, the element and its
preceding silent interval are the unit of song in the Zebra Finch.
Each male sings a number of different elements in a set sequence and
together these constitute the song-phrase (Hall 1962; Immelmann 1965a,
1969) ('song-unit', Price 1979; 'motif', Sossinka and Bohner 1980; 'song'
Cynx et al. 1990). Within one performance, males may repeat the song-
phrase a number of times to form a song ('bout' Price 1979; 'strophe'
Sossinka and Bohner 1980), the first phrase of which is usually preceded
by several identical elements, which form the song introduction. Sorne of
these introductory elements may be incorporated into the phrase itself
(e.g. Figure 10.6b). Intervals of approximately five seconds separate the
216 THE ZEBRA FINCH
(a) (b)
~i \
t~: : '
,
:::
kHz
(e)
8
0.5 s
4
·¡P
end of one song and the start of the next during which the male resumes
a normal perching posture. The song-phrase is the natural unit of investi-
gation of Zebra Finch singing and is used by most researchers, rather
than song, since the number of song-phrases sung per song depends
on motivatíon and varíes according to whether the song is directed or
undirected (see below).
Song-phrase macrostructure (Table 10.2)
Considerable variation exists in the gross features of the song-phrase
among indíviduals, populations and regions (Zann 1993a,b); variation
also exists between domesticated and wild forms (Slater and Clayton
1991; Zann 1993a), and between the two subspecies (Clayton et al.
1991). In wild Australian Zebra Finches, the number of elements sung
10 Vocalisations 217
Table 10.2 Comparison of song-phrases from Australian and Lesser Sundas Zebra
Finches showing medians and interquartile ranges; after Clayton et al. (1991) and
Zann (1993a).
Fig. 10.8 Fourteen types of song-elements sung by wild Australian Zebra Finches.
Horizontallines show 1 kHz intervals. (After Zann 1993b.)
(3) the Distance-call Element, whose structure matched that of the Dis-
tance Call, constituted 16% of all song-elements, and was sung by
87% of males; in 50% of males the Distance-call Element was the
same version used as the Distance Call, but much softer.
Immelmann (1962a) was the first to identify the presence of call-like
elements in the song phrase, and Price (1979) later distinguished three
types, short, medium and long that correspond respectively, with the Tet,
Stack and Distance-call Elements.
The remaining 33% of song-elements were unique to the song, having
no direct structural resemblance to any calls (except for the Introductory
Diad), and were composed of eight different types, of which the High
Element formed 10%, and occurred in 67% of males. This distinctive
element was by far the briefest (median duration 8 ms) and highest-
pitched (median fundamental frequency 6.4 kHz) sound given by Zebra
10 Vocalisations 219
Finches, and was the only sound that did not have numerous harmonics
(see Figures 10.6 and 10.8). Its fundamental frequency was five times
greater that of the next-highest-pitched element in the song and Williams
et al. (1989) believe its structure has arisen from suppression of lower
harmonics. By contrast, the remaining eight types of song-elements con-
tained various combinations of unstructured noise and harmonic spectra,
arid their occurrence was variable among males. Production of these
compound elements was less stereotyped than other elements and varied
within and among individuals; consequent1y, unambiguous classification
is more difficult. Criteria for identifying these less-diagnostic elements,
are given in Zann (1993a).
The number of element types sung by Lesser Sundas Zebra Finches is
unknown. No comparisons have been established between the caH reper-
toire, which itself is still largely undescribed, and the song elements,
although some males appear to produce elements that faint1y resemble
Distance Calls.
Sequence of elements
Each male sings its song-elements in a fixed order and the pattern is
fairly consistent among' individuals. With Australian males it is possible
to recognise three sections to the stereotyped song-phrase: a start, middle
and an end. The first section consists most1y of Introductory Elements; a
single High Element is the centrepiece of the middle section, and is pre-
ceded and foHowed by compound elements; the end section consists
most1y of the Stack Element and the Distant-call Element, in that order
(e.g. Figure 10.6b), but not always (e.g. Figure 10.6a). The Distance-call
Element is much louder than the rest (Zann 1993a), and the most resis-
tant to attenuation, so that it is often the only element heard clearly from
a distance, or in strong wind. With its position at the end of the song-
phrase the Distance-call Element serves as a 'coda' or 'punctuation' that
separates one phrase from the next.
The organisation of the song-phrase into three sections of elements
appears to have sorne functional basis. When production of song in
domesticated Zebra Finches was experimentally disrupted, Williams and
McKibben (1992) found that elements disappeared, and were later
recovered, not singly, but in 'chunks'. Furthermore, during song learning
there was a tendency to learn elements from the song tutor, or model, in
chunks. These chunks correspond to the sections identified in the song-
phrase of wild Zebra Finches
Evolution of the song-phrase
The sequence of elements in the song-phrase of the Australian subspecies
shows similarities with the series of calls given around take-off, so
conceivably, the ancestral song may have originated from such a
sequence (see aboye). Unfortunately good sonograms of the take-off calls
220 THE ZEBRA FINCH
are difficult to obtain beca use the exertions involved in lift-off distort
call structure (Zann 1984). In general, the same sequence occurs in the
song-phrase, but inserted between the Introductory Elements and the
Stacks is the middle section of non-call-like song-elements, namely those
that are original and exclusive to the song, and include the High Element
and its neighbouring compound elements (Figure 10.9).
The Introductory, Stack and Distance-call Elements may have been
'borrowed' from their countcrparts in the call repertoire, but the origins
of the non-call-·like eJements are not obvious. Nevertheless, a11 ten ele-
ments havc probably been derived from the Distance Call through sys
tematic repetitíon, modification and omission of the tonal and noise
components of thc call (Zann 1993a). Thus, in this thcoretical scheme
the Distance Call has a dual role in the evolution of the elements of the
song-phrase-it not only provides a sound that is used unmodified in the
end section of the song, but it also serves as a source from which the
non-call-like elements in the middle scction have bcen derived. One can
only speculate on how such an arrangement might have ariscn. One pos
sibility is that the original ancestral song consisted of three unlearned
calIs given at take-off, after which the fcmale-like Distance Call was sub-
jected to sexual differentiation during which it carne under control of
higher vocal centres. Subsequcntly, during its ontogeny the Distance Cal!
beca me open to auditory input from an external model, perhaps from
the father, and the derived elements were a by-product of this process.
The last step would require aH the song elements to be copied from the
external model in one process, thus emancipating the entire song-phrase
from the lower vocal control centre.
The changes that have produced the song phrase of the Australian
Zebra Finch may be of fairly recent evolutionary origino The incorpora-o
Fig. 10.9 Schema showing sections of the song-phrase of the Australian Zebra
Finch and the possible cvolutionary sourcc of the elcments from the call repertoire. 1,
Introductory Element; L-N, Ladder-noise Element; H, High Element, N-S T,
Noise-strueture-Tone Element; S, Stack Element; D-c, Distance-eall Element; Te,
Tet Call, DC, Distance Call; SC, Stack Callo
10 Vocalisations 221
tion of call-like elements into the song does not occur with the Lesser
Sundas subspecies. Conceivably, the call-like elements have been lost
since geographic isolation, but it is more likely that they only evolved in
the Australian subspecies after isolation. Absence of sexually dimorphic
Distance Calls, the source of non-call-like elements, suggests they never
existed in the Lesser Sundas subspecies. Finally, none of the three species
of Poephila have sexually dimorphic distance calls, nor do they incorpo-
rate them into their songs (Zann 1976b).
Song variation
While there are fairly clear rules of organisation of the song-phrase in
wild Australian Zebra Finches, each individual sings a unique, stereo-
typed phrase. Recording of songs fram the same individual up to one
year apart show phrases identical in element number, structure and
sequence, and singing duration (Zann 1990). Similar stability exists in
domesticated Zebra Finches (Nordeen and Nordeen 1992). Thus, there is
the potential for the song-phrase to serve as a permanent individual sig-
nature.
Geographic variation in the song-phrases of 402 males from 33
colonies in two distant parts of Australia was investigated using canoni-
cal discriminant classification of 31 variables (Zann 1993b). Song-
phrases fram southeastern Australia were clearly distinguishable fram
those of central Australia. While there were no differences in the dura-
tion of the phrase, there were significantly more elements in those fram
central Australia (7 vs. 6); moreover, they were sung more rapidly (8.5
elements per s vs. 7.4) and had a higher fundamental frequency. The
composition and sequence of non-call-like elements in the middle section
of the phrase also differed significantly between the two geographic
zones. When analysis was restricted to southeastern Australia, variation
in the middle section again distinguished song-phrases fram different
colonies, and their position in the classification was more or less propor-
tional to the geographic distance separating them and the estimated
likelihood of dispersal. When analysis was focused down to a few
colonies, significant variation in the middle section of the song-phrase
was found from one year to the next and depended on the origin of the
large influx of immigrants (Chapter 8).
In summary, variation in the song-phrase of the Australian Zebra
Finch occurs principally in the non-call-like elements that constitute the
middle section. Whereas the High Element is almost a permanent fixture,
the occurrence of the nine compound elements is quite variable, and
these form the most labile part of the phrase. By contrast, the call-like
elements, at the start and end sections are stereotyped. Similarly, the
macrastructural features, (phrase duration, element number, speed of
singing) are not highly variable across colonies, and so must be more
rigidly specified. This was confirmed by (Clayton 1990b) in hybridisation
222 THE ZEBRA FINCH
during Directed Song except for the 'angular head' (Figure 9.5c). No
differences are detected in the head movements during both types of
singing. As each phrase is sung the bill opens and closes once or twice
and the head pivots to and fro to one side, swinging about 45° from the
forward position; it pivots to the other side for the next phrase.
Intensity differences in the acoustic and visual components of singing
have led a number of authors to conclude that Undirected Song and
Directed Song have a common control mechanism in which Undirected
Song is simply the sexual display with the lowest threshold, whereas
Directed Song is the one with the highest threshold (Morris 1957;
Kunkel 1959; Immelmann 1962a; Güttinger 1970). In contrast, Caryl
(1981) concluded that Directed and Undirected Songs did not share a
common control mechanism because he found no correlation between
rates of Directed Song and Undirected Song, and the sexual stimuli
important for Directed Song were not important for Undirected Songo
Both types of song are strongly influenced by gonadal hormones, con-
sequently there is a reduction in singing if males are castrated, and a
restoration of singing if there is replacement with gonadal androgens
(Prove 1974; Arnold 1975a; Harding et al. 1983). Frequency of songs
was positive1y correlated with Ieve1s of circulating testosterone, but
threshold concentrations were significantly smaller for Undirected Songs
(Prove 1978; Prove and Immelmann 1982). Nevertheless, analysis of
how androgens affect the production of Directed and Undirected Songs
indica tes that they have separate controls and are not simply at different
points on the same unitary song system. Walters et al. (1991) discovered
that Directed Singing is oestrogen dependent and causes songs to be sung
faster and to be directed at females, whereas Undirected Song is more
224 THE ZEBRA FINCH
androgen dependent and is either independent of oestrogen entirely, or
extremely sensitive to smalI quantities. How these hormones affect the
control circuitry in the brain is yet to be determined.
Functions of song
As with most estrildines, song appears to have no territorial function in
Zebra Finches (Morris 1954; Kunkel 1959; Hall 1962; Immelmann
1962a, 1965a). They possess no territory and males take no aggressive
action towards singing males, nor do singing birds behave aggressively
before or after singing (Dunn 1994).
Function of Directed Song
It was generally assumed from its context in the courtship display that
Directed Song was a pre-copulatory signal that elicited female solicita-
tion (Morris 1954), however, it was not until experiments by CIayton
and Prove (1989) that this assumption was confirmed. Semi-domesti-·
cated females of both subspecies of Zebra Finches and Bengalese Finches
were implanted with oestradiol to heighten their sexual response and
then placed alone in sound-shielded boxes where they were subjected to
songs played back through a loudspeaker. Female responsiveness
towards the songs was measured by the number of copulatory solicita-
tions (Tail-quivers). Females responded strongest to songs of conspecific
males, preferring them to those of heterospecifics; moreover, the Zebra
Finch females preferred the song of theÍr own subspecies, thus con-
firming the hypothesis that Directed Song on its own serves as a means
of recognisíng species and subspecies. Cross-fostering experiments
between subspecies showed that this preference is learned from exposure
to the songs of their fathers before 35 days of age (Clayton 1990b).
Females also discriminated among songs within subspecies. They pre-
ferred long song-phrases over shorter ones, that is, phrases with more
elements rather than phrases with fewer elements. Presumably, a male
with a long complex song-phrase would not only elicit a copulation
solicitation more rapidly from his mate, but would more be successful
with extra-pair females as well, and probably be preferred at pair forma-o
tion over males with simpler songs. In choice-tests females are attracted
to males that sing frequently (ten Cate and Mug 1984; Collins et al.
1994). Provisional results from other captive studies suggest that females
have a higher reproductive success if they can paír with males whose
songs they prefer (Williams et al. 1993). Both studies support the
hypothesis that sexual selection plays an important role in the evolution
of estrildine songo
Directed Song is also used for individual and kin recognition, a suspi-
cion long held by Immelmann (1968, 1969) on the basis of the large
individual variation in song structure in domesticated and wild-caught
birds. Playback experiments with domesticated Australian Zebra Pinches
10 Vocalisations 225
showed that females could recognise songs of their mates and preferred
them to similar, yet familiar songs of other males (Miller 1979a). Further
experiments showed that daughters also preferred the songs of their
fathers over that of other males even after many months of separation
(Miller 1979b). The ability to memorise the father's song for the purpose
of recognition and discrimination first occurs in both males and females
between day 25 and day 35, and the ability extends to a least six months
of age (Clayton 1988a). Finally, it is likely that adult females quickly
learn to recognise the song of new partners.
Function of Undirected Song
In contrast to Directed Song, the functions of Undirected Song are less
obvious. Immediate overt responses are not observed among potential
recipients of Undirected Song in captive and wild Zebra Finches.
Moreover, the absence of detailed field studies and experiments has,
until recently, limited knowledge of its function to mere speculation.
Immelmann (1968) for example, concluded that it was basically
functionless and resembled subsong and was purely an indication of a
'very tranquil mood'. Recently, however, Dunn (1994) focused investiga-
tions specifically on the contexts of Undirected Song in the wild and
conducted experiments on both wild and captive birds. He concluded
that Undirected Song not only has costs in terms of energy expended and
risks taken but that it has multiple functions. Undirected Song occurred
at two main locations in the Danaher colony-around the nest during
the breeding season and at flock feeding sites. Undirected Song occurred
in the flock throughout the day with slightly more males singing in the
morning than the afternoon, but within males rates were similar. Males
sang Undirected Songs in the flock throughout the year but a greater pro-
portion sang during the non-breeding sea son than during the breeding
season, but the individual rate of singing was no higher. Although testos-
terone levels fell in the non-breeding season this did not affect the rate of
Undirected Song in the flock. When female partners were experimentally
removed the rate of Undirected Song increased significantly and fell
again when the partner was returned, or when the male re-paired. The
finding suggests that the female partner inhibits, but does not stop, male
Undirected Song performance in the flock, and that males use Undirected
Song to attract females for pairing purposes or to advertise their avail-
ability for extra-pair copulations; they may also advertise their quality so
that they might rapidly re-pair should they lose their current partner.
During the breeding season, Dunn (1994) found that Undirected Song
occurred most frequently during the nest-building and early egg-Iaying
stage. Typically, the male sang when the female was in the nest and he
had just left her. If the male sang immediately the female entered the nest
it significantly increased the time she spent inside. This enhanced the
male's ability to guard the female from extra-pair matings during her
226 THE ZEBRA FfNCH
fertile period, and also had the advantage of preventing brood parasites
from dumping eggs, or other pairs from taking over the nest. Thus, a col-
lateral function of Undirected Song at the nest site is mate- and nest-
guarding. Undirected Song tends to decline during incubation but
increases after young hatch. Ten Cate (1982) found that Undirected Song
increases steeply in domesticated Zebra Finch fathers from day 8 to day
24, after which singing occurs at sustained levels.
Development of song
Like most song birds Zebra Finches learn their songs early in life. Immel-
mann (1965c, 1967, 1969) made this discovery in pioneering experi-
ments in which he manipuIated the early social life of young males by
exposing them to different auditory experiences extending fram the egg
stage up until sexual maturity. Sorne males he crass-fostered to Bengalese
Finches in sound-shielded boxes where they were isolated fram aH
sounds made by Zebra Finches. He prevented others fram hearing any
songs at all until maturity, by either hand-raising them or by having
females raise them unaided. Finally, he had others cross-fostered to Ben--
gaIese Finches, where they couId he al' and see other Zebra Finches. His
results can be summarised as follows:
(1) Zebra Pinches must learn the details of their song-phrase since only a
raugh version exists without learning (see below);
(2) they can learn the complete song of the Bengalese Finch foster-father
which completely masks the unlearned Zebra Finch framework;
(3) elements of the song can be learned and memorised as early as day
25, about the time they start performing subsong, but after the fuI!
song is performed, around day 80, new elements cannot be learned--
a sensitive phase for song-Iearning exists in the first three months of
life;
(4) Zebra Pinches learn their songs fram the male that feeds them, so it
is usually the father or foster-father who serves as their song tutor;
(5) if there is no father 01' foster-father, young males show a bias
towards learning the songs of Zebra Finches in preference to those of
other species.
Immelmann (1969) concluded that wild Zebra Finches would be likely to
learn songs of their fathers, and an early end to the sensitive phase was
necessary in order to prevent accidental learning from heteraspecific
estrildines that they might encounter.
In due course this fascinating investigation inspired a long series of
intriguing follow-up experiments by other researchers, in particular, by
P. J. B. Slater and co-workers in the last decade, who used song learning
in domesticated Zebra Finches as a model for teasing apart the subtle
interactions involved in the development of behaviour. A detailed review
10 V ocalisatim~s 227
of pragress made in understanding Zebra Finch song development fram
Immelmann's initial work up until 1988 can be found in Slater et al.
(1988). An early assumption of Slater's appraach to song learning was
that wild Zebra Finches would be unlikely to have their father available
as a song tutor after 35 days of age because bonds between them would
be broken at about this time. It was reasoned that wild young become
independent around 35 days (Immelmann 1962a) and the re-nesting
father would tend to drive them away. Consequently, his group initially
focused on how young males chose a song tutor other than the father
and what factors delimit the sensitive phase. My studies at the Danaher
colony show that although nutritional independence is almost complete
by 35 days, roosting bonds can continue with the parents until around
50 days and association in feeding flocks is likely beyond that age, espe-
cially when one considers the abilities of kin to recognise one another
(Chapter 9). AIso, no observations have been made of the father driving
away his young. Consequently, contact with the father and his song is
highly likely beyond 35 days providing opportunity for the father to
serve as tutor in wild Zebra Finches.
Physical development of singing
The first subsong begins in wild birds between 28 and 35 days of age
when still in the company of their parents and siblings, but it is so soft
that when one observes fram as little as a few metres distant, the singing
posture is the only clue to its performance. The upright stance, pivoting of
the head and open and closing of the bill fo11ow the adult display except
that the plumage is uniformly fluffed. Up close, quiet bursts of toneles s
sound are heard which are punctuated with irregular squeal-like warbles.
Within a week or so subsong gets louder and is often preceded and inter-
rupted by a series of Distance Ca11s, during which the plumage sleeks
again. The calls get softer before the transition to subsong and sorne are
incorporated into the rambling unstructured performance. After day 35,
males are often seen singing alone on the tops of bushes. In captivity,
there is great individual variation in the onset and frequency of subsong.
Arnold (1975b) detailed the changes in song structure fram 40 days of
age to adulthood. At first elements are poorly formed and highly variable
in structure and sequence, and no song-phrases are discernible-the song
has a rambling or babbling quality. By day 50, element variability is
reduced and there is a vague resemblance to adult song in element form
and organisation, and by day 60 almost a11 elements of the final song are
recognisable, but their praduction is not as stereotyped as that of adults.
A sequence of elements is recognisable by day 60 although slight changes
may subsequently occur. By 90 days of age the song contains elements
that are sung with the stereotyped form and sequence, typical of adults
more than four months of age, although a sma11 proportion of elements
can change slightly after this age (Morrison and Nottebohm 1993; Slatcr
228 THE ZEBRA FINCH
et al. 1993). The song-phrase shortens slightly after day 90 due to a
deerease in the interval between the e1ements; in addition, the number of
song-phrases per song inereases. In summary, Zebra Fineh song develop
ment is like that of most species of song birds where three phases are dis-
cernible: (a) the rambling subsong phase between 25 and 50 days that
bears little resemblanee to the adult song; (b) a plastic song phase
between 50 and 80 days in whieh sorne features of adult song are present
and during which the elements of the adult song gradually 'crystallise'
out to form (e) the final song phase. It is not elear whether Zebra Finehes
overproduee song elements at the onset of plastie song and gradually
lose all but those reprodueed in the adult song via a proeess of seleetive
attrition (Marler and Peters 1982), or the final version erystallises out
sorne other way.
Motor development of song in Zebra Finehes is temporally eompressed
in eomparison with other species of song birds that ha ve been studied.
Consequently, the two stages neeessary for development of song, in the
Zebra Finch (namely, memorisation of the aeoustic structure of the
tutor's song and the correet development of the sensory-motor integra
tion that permits reproduetion of that song strueture) overlap in time. In
other species of song birds, such as the Swamp Sparrow Melospiza geor-
giana, the two processes are temporal1y discrete (Marler and Peters 1982).
Songs without tutoring
In his preliminary experiments Immclmann (1967, 1969) showed that
males deprived of the opportunity to hear other adult finches from the
nestling stage to maturity sang aberrant, self-improvised elements. Nev
ertheless, these untutored songs still retained the speeies-typieal temporal
patterning that formed the macrostructural organisation of song-phrases
and song-bouts; in other words, the contents were abnormal, but the
packaging still fairly normal. This was subsequendy eonfirmed by Price
(1979), Clayton (1990b) and Morrison and Nottebohm (1993). Price
coneluded that the conserva ti ve macrostructural organisation of song in
the Zebra Finch is due to a 'neuromotor eonstraint'. Strangely, not al!
Zebra Finehes showed this constraint; the songs of birds visually and
aeoustieally isolated by EaJes (1985) after day 35 displayed no phrasing
at aH, and had many repeated elements. Coneeivably, the early exposure
to the father's song in Eales' study somehow disrupted the unlearned
framework, or there may have been important differenees in breeding
stock from whieh experimental animals were sampled.
Element structure is abnormal and variable in untutored males, and
mostly consists of non-eall-like elements (Priee 1979). Untutored elc-
ments are abnormally high in frequeney and long in duration, ano
inelude many with upward infleetions, and others that are eliek-líke
(Morrison and Nottebohm 1993; Williams et al. 1993). If untutored
males are held in groups they tend to learn from eaeh other (Slater et al.
10 Vocalisations 229
1993). Call-like elements are not found in songs of isolates although
their Tet and Stack Calls develop normally. It appears that males cannot
form song-elements out of their own calls, but must learn them from
other individuals. Wild Zebra Finches incorporate the father's Distance-
call Element or his Distance Call into their song-phrase (Zann 1990),
and possibly his Tet and Stack Calls as well. Similarly, untutored males
raised in isolation with their mothers, or foster-mothers, will learn the
female Distance Call and incorporate it into their aberrant songs (Immel-
mann 1967, 1969; Price 1979, Eales 1987a); a phenomenon also found
in Bengalese Pinches (Clayton 1987b). Price (1979) concluded that incor-
poration of developmentally conservative calls into the song-phrase was
a means by which element formation could be rigidly specified, and so
preserve the species song structure from accumulating learning errors
that might occur with the developmentally labile non-call-like elements.
The developmentally conservative Tet and Stack Calls could serve this
purpose, but not the Distance Call, whose noise component must be
learned, and whose absence from many Distance-call Elements of domes-
ticated Zebra Pinches is a manifestation of its developmentallability and
high susceptibility to copying errors.
Sensitive phase for song learning
Both domesticated and free-living Zebra Finches mainly learn their songs
in the second month of life, specifically between days 35 and 65. This
was first shown by Eales (1985) who removed domesticated young males
from their father's cage at day 35, 50 or 65, and either held them in
visual and acoustic isolation from all adults until day 120, or held them
in a cage next to a pair of adults Zebra Finches, the male of which served
as a new song tutor. Those given the new tutor at day 35 learned all
their song from him, while those given the new tutor at 65 days kept the
father's song and learned nothing; those switched at 50 days learned a
hybrid songo In the isolated group the proportion of song-elements
learned from the father was in direct proportion to the time spent in his
cage, and by day 65 all his song had been learned. However, those iso-
lated at 35 days sang aberrant, amorphous songs, but when given access
at six months of age to normal male Zebra Finches in an aviary, they
learned a stereotyped, species-typical song-phrase. Strangely, when this
last experiment was replicated by Slater et al. (1993) the majority
showed signs of learning from the father and only modified theirs songs
slightly when exposed to adult male tutors after 120 days. When Eales
(1985) had males raised alone by their mothers they sang abnormal,
unstable songs, but subsequent exposure to normal male tutors at 65
days of age showed that they were still capable of learning his songo
When they did, it replaced the earlier version based on female ca lIs (Eales
1987a); this also happens with Bengalese Pinches (Clayton 1987b).
These results indicate that domesticated Zebra Finches mostly learn
230 THE ZEBRA FINCH
their song between 35 to 65 days of age, but if no acceptable tutor is
available during that interval the ability to learn can be extended until
one comes to hand, if not, songs learned before day 35 can be used. The
important point is that it is experience, in combination with age, that
determines the end of the sensitive phase. On the basis of their findings
on untutored adult males, where song learning could be extended well
into adulthood, Morrison and Nottebohm (1993) concluded that the
closing of the sensitive phase probably results from social interactions
with the tutor which ¡ead to the acquisition of a 'stable motor memory
of song'. Experiences such as singing a stereotyped song, hearing and
copying elements from others, and development of particular types of
song were considered to be less important factors in the closing of the
sensitive phase. Therefore, Zebra Finches are clearly not 'open-ended'
learners like canaries that can learn a new song each breeding sea son
(Nottebohm 1993), but they are not strictly 'age-specific' learners either.
Most wild Zebra Finches will also learn their song during the 35-65
day sensitive phase. This was shown at the Danaher colony by compar-
ing songs of fathers and their sonso The sons, who were caught during
the second month of life and caged in an aviary within the colony were,
in effect, social1y isolated fram the father (Zann 1990). There was a
significant positive correlation between the time of exposure to t:he
father outside the aviary and the proportion of elements shared. Those
sons not caged until 65 days of age showed a strong match with
the father's song, whereas those caged soon after day 35 showed no
resemblance at aH.
Learning before day 35
By swapping nestlings between nests of different males, Arnold (197 5b)
showed that no learning occurred before fledging. After fledging, the
father's song is completely memorised by the son before day 35, but
whether he sings it or not depends on what happens during the second
sensitive phase. If he receives inadequate tutoring during the second sen-
sitive phase, or it is disrupted to sorne extent so that there is a mismatch
with the visual and auditory stimulation he received before day 35 he
wiU produce the song learned before day 35 from the primary tutor. For
example, if deprived of auditory, visual and physical contact with other
Zebra Finches he will sing the song learned before day 35 just as
completely as controls kept continuously with the father for 100 days
(B6hner 1990). Similarly, if he cannot see the tutor, or cannot interact
vocally with him (Eales 1989), or is exposed to a tutor of a different
species to his rearing father (Clayton 1987c), or to a different morph to
that of his rearing father (Mann et al. 1991, Slater and Mann 1991), or
exposed to two tutors of different species, either successively (Clayton
1987d), or simultaneously (Clayton 1988b), he wiU sing the father's song
heard before 35 days of age. Finally, if the acoustic structure of songs
10 Vocalisations 231
and Distance Calls of the new tutor diverge too strongly from those of
the father they will be rejected (Slater and Jones 1995).
Exactly how Zebra Finches learn their father's song between fledging
and day 35 is unknown and requires investigation.
Learning from the father
Immelmann (1967, 1969) found that domesticated male Zebra Pinches
learned the song of their rearing fathers even when cross-fostered to Ben-
galese Finches and able to hear, see and interact with conspecifics during
the sensitive phase. Other semi-naturalistic aviary studies found that sons
copy their father's song in preference to other Zebra Pinches (Arnold
1975a; Schwab 1986; Williams et al. 1993), but not in every case
(50%-Mann and Slater 1995), and in one study sons showed no prefer-
ence at all for the father's song (Williams 1990).
In an attempt to determine if wild birds sing their father's song, 1
compared song-phrases of 40 sons with thoseof their 20 rearing fathers
during a three-year study at the Danaher colony (Zann 1990). In order
to detect evidence of copying, sonograms of song-phrases each of the
two individuals in question were examined element by element for
matching, and a per centage score of matched elements compiled. Con-
ceivably, two phrases could have elements matching by chance, that is,
without any copying or transmission from one individual to the other.
To determine what level of matching occurred by chance in this popu-
lation 1 compared the song-elements of 57 dyads chosen at random
from 55 unrelated mature adults, and found that a median of 20% of
elements matched, with a range of 0-57%. Next, 1 determined how
high a matching score had to be for it to be statistically unlikely to have
been drawn from the population of chance matches, and found that the
probability of matching scores of ~ 54% arising by chance was les s
than 2.5%. Therefore, scores greater than this were unlikely to occur
by chance and may have arisen by copying. 1 estimated that sorne 61 %
of sons matched the songs of their respective fathers and presumably
copied from them. Sorne sons matched their father's song in every ele-
ment, whereas others, even in the same brood, showed poor matches,
sorne not exceeding chance levels (e.g. Figure 10.11, son (b) vs. son (c)).
In one family, it was possible to compare song phrases across three gen-
erations, a grandfather, father and four grandsons. Matching scores
reached criterion in all but two grandsons, one of whom was confined
to the aviary during his sensitive phase and effectivelY excluded from
learning the father's song (see Figure 3 in Zann (1990)).
Slater and Mann (1990) have argued that it is conceivable that a son
did not actually learn from the father himself, but from a tutor that hap-
pened to sing like the father. Unless this tutor was a relative of the
father's it is unlikely that such an individual would be present in the
colony since the probability of a chance matching with the father is only
(a) Father
F A B e D E F
A B e D E F A B e 12 E
N
I
.o.t.
>.
()
4
e
Q)
::J
CT
Q)
U:
O
+ A + + D + A B D E E
0.55
Fig.10.11 Song-phrases of a father (a) and his four sons from two broods; sons (b)
and (e) were from one brood and (d) and (e) from another. Two song phrases of the
father are shown and one of eaeh son. Elements of sons that match those of the
father have the same letter; underlined letters are good, but not perfeet matehes, and
a + sign indieates that the element has no matehes. The pereentage of a son's ele-
ments that matehed those of the father are: (b) 100%, (e) 33%, (d) 83%, and (e)
92%. (After Zann 1990.)
10 Voealisations 233
2.5%. High song diversity at Danaher is probably due to the large influx
of immigrants each year (Chapter 8). Moreover, the probability that sons
in successive braods and in successive generations would encounter this
same individual that happened by chance to sing the family song and
learn from him are remote. Finally, I showed that Distance-call Elements
and Distance Calls match between fathers and sons and sorne matches
crass four generations-further evidence for sons learning fram the
father tutor. Therefore, the most parsimonious explanation is that sons
whose songs matched those of their fathers actually copied from them.
This is consistent with much of the laboratory work. What is difficult to
explain is why about 40% of sons learned fram other free-flying males,
rather than their own father despite his availability as a song tutor. This
can even happen in the one braod where one sib will accurately copy the
father and another will only show chance levels of matching, and pre-
sumably copied someone else. This presents a new problem that can only
be answered in the laboratory, namely, how are song tutors chosen?
How is the father eh osen as a tutor?
In the laboratory a young male willlearn the father's song in preference
to that of another male, praviding that the father is present during the
second sensitive phase. This will occur whether the youngster is reared
with the father for the first 35 days of life (B6hner 1990) or reared by a
single mother and only encountering the father for the first time after day
35 (Eales 1987b). Experiments indicate that a range of cues are used to
recognise the father.
Immelmann (1967, 1969) noticed that Zebra Finches tend to learn
songs fram those males that feed them, and this individual is usually the
father in domesticated birds (and exclusively so in wild birds). However,
in sorne domesticated stock, males other than the father will feed young
and serve as song models (Williams 1990). Thus, the father-son nutri-
tional bond may be one way a son recognises the father as a suitable song
tutor; the occurrence of Undirected Song before and after feeding of young
may be important for the learning pracess. Since pravisioning care ceases
before the main sensitive phase begins, this might not be the proximate
mechanism that leads a son to chose the father as a model after day 35.
Both B6hner (1983) and Eales (1987b) noticed that fathers in cages
behave aggressively to their young, and since aggression is believed to
have an important influence on sexual imprinting in the Zebra Finch (ten
Cate 1984) they thought that males may choose to copy the song of the
most aggressive male they encounter. Indeed, when two non-fathers were
offered as tutors Clayton (1987a) found that the one most aggressive
towards the young male was the one he chose to learn from. However,
recent experiments indicate a preference to learn fram males aggressive to
other individuals, and not necessarily the one aggressive to the young maJe
seeking a song tutor (P.J.B. Slater, pers. comm.). In a small breeding cagc
234 THE ZEBRA FINCH
sueh, but probabIy have a preferenee for tutors that have Zebra Fineh
ca lIs and Zebra Fineh song maerostrueture (CIayton 1989a).
The need for visual and vocal interaetions between young male Zebra
Finehes and their song tutor was impressively demonstrated by Eales
(1989) who exposed males to (a) tutors they eouId see and with whom
they could interaet voeal1y, (b) tutors they could not see but couId inter
aet with voeally; and (e) tutors they eould hear, but not see or interact
with vocally. Males in group (a) all learned from the tutor, whiJe nonc
did so in group (e). Nevertheless, in group (b) half learned from the tutor
they could not see which suggests that the visual components of the song
tutor are not always essential for song learning. This aspect has been fol··
Iowed up recently by Adret (1993) who placed young male Zebra
Finches in a Skinner box and trained them to peck at a key that resulted
in the broadcast of a Zebra Finch song from a tape via a loudspeaker
placed inside the eage. The test male eould hear, but not see a control
bird, a male sibling companion that could also hear the taped song, but
had no switches to control its production. Test males not only found the
song reinforcing and worked the key in order to hear the taped song, but
they also memorised it and sang it at maturity. In contrast, the control
males barely learned any elements of the songo Adret conduded that
operant eonditioning with song as a reward strongly influences song
learning, that is, the young male is not sorne passive 'sponge' soaking up
the song of the selected tutor, rather it exercises sorne control over the
learning proeess. Exactly how this might work in the natural eontext is
unknown. In a similar Skinner box experiment ten Cate (1991b) also
conditioned test males to song playback, but was unable to demonstrate
song learning, perhaps beca use the operant response (perehing) was less
appropriate than that in Adret's study (1993)
Ten Cate (1986a) noticed that foster-reared male Zebra Finches often
sat very close to the singing Bengalese Finch foster-father and appeared
to 'peer' at, or 'listen' intendy to him. Listening such as this is extremely
rare in normally raised Zebra Finches but is common among young man-
nikins (Morris 1958). Ten Cate believes that listening is a means of
attending earefully to the song performance in order to copy it, and may
be essential when the song to be learned is soft, complex and not sung
very often, which is often the case with mannikins.
Overview
Both field and laboratory findings suggest that sons are indined to learn
their fathers' songs. There appears to be an overlap in the meehanisms
that promote learning of the father's song, and additional mechanisms to
ensure that they at least learn the songs of Zebra Finches, rather than
those of heterospecifics. AH details of the father's song are learned in the
first sensitive phase during early post-fledging and this is the primary
backup system, an insurance against failures to learn the conspecific song
10 Vocalisations 237
during the main sensitive phase between 35 and 65 days of age. During
the second month of life, field data suggest that young Zebra Finches still
have tight social bonds with the parents, and the early end to the sensi-
tive phase fosters the learning of song from the father. Furthermore, if an
incorrect song is learned, the sensitive phase remains open until a satis-
factory tutor is encountered and a normal song learned. During the
35-65 day interval, the young male is inclined to learn from the individ-
ual that should be his father; moreover, the fact that he prefers to learn
from a tutor that sings like his father appears to be another means of
guiding his song towards the paternal version. Finally, imprinting on the
siblings and/or parent is a means of ensuring, at least, that a Zebra Finch
will be the tutor.
Bohner (1990) proposed a two-step learning process. The first stage
occurs between fledging and day 35 during which aH the details of the
father's song are memorised. After day 35, the onset of the second stage
of learning, no new song will be acquired if the father continues to be
present during the main sensitive phase, but if he is absent a new tutor is
chosen. This new tutor is not chosen at random, but is guided by the
visual and vocal experiences gained before day 35.
Given the redundancies built into the song learning process that
appear to promote learning ofthe father's song in preference to that of
other males it is puzzling that only about two thirds of wild males
appear to learn the father's song at all, and sorne of their copies are not
very accurate. Even in cage studies, where only a few tutors are present,
inaccurate copying is a common feature of song learning, although one
benefit is that it enables individuals to develop unique versions that can
be used for individual recognition (Bohner 1983). Variability in the
learning process in the wild, especially among siblings, is unexplained,
but recent findings by Jones and Slater (1993) raise the possibility that
females, other than the mother, also influence what song a male devel-
ops. In this way males arrive at a song that females find attractive yet is
not necessarily the same as their fathers'. This situation may be compara-
ble to 'action-based' learning found in sorne other species of songbirds
(Marler and Nelson 1993).
In his initial study, Immelmann (1967, 1969) found that the Bengalese
Finch and the African Silverbill Euodice cantans also learned their songs
during an early sensitive phase and could learn from heterospecifics in a
manner similar to that of the Zebra Finch. Subsequent investigations by
Clayton (1987b) showed that song development in the Zebra Finch and
Bengalese Finch is almost identical in all details, except that the latter
does not indude caH-like elements into its phrase. Unfortunately, few
studies of song development have been made on other species of
estrildines to determine whether all estrildines possess a short sensitive
phase early in life or whether it only occurs in those that have particular
ecological adaptations. In considering the biological advantage of a short
238 THE ZEBRA FINCH
and early sensitive phase in Zebra Finches, Immelmann speculated that it
was a means of ensuring that only the conspecific song would be learned,
thus precluding accidental learning from any heterospecific a young male
might encounter in mixed breeding colonies or in mixed feeding flocks
later in life. Too little is known about the biology of free-living Silverbills
and the wild ancestors of Bengalese Finches to determine if this explana-
tion also applies to them. Moreover, with Zebra Finches, the risk of accÍ-
dentallearning from heterospecifics appears low under present ecological
conditions, since mixed colonies and mixed feeding flocks are rare
(Chapter 4). Conceivably, during periods of intense aridification, Zebra
Finches may have been forced to move out of much of the interior of
Australia to seek refuge in areas where close living with other species of
estrildines occurred and so heightened risk of learning incorrect songs.
The provision of numerous, widely dispersed, artificial water points since
European settlement may have limited the extent to which Zebra Finches
are now forced to disperse, and consequently their exposure to other
estrildines has become limited in recent times. Unfortunately, too little is
known of the comparative breeding biology and song development of
wild estrildines to proceed beyond the speculation stage.
Control of song
Can aries and domesticated Australian Zebra Finches are the two species
of songbirds used as experimental models for investigations of brain
function, in particular, the neural basis of learning and mechanisms lead-
ing to sexual differentiation in the brain. lnvestigations of the vocal con·-
trol system in these species offer 'a unique opportunity to study the
interplay of hormonal, environmental, and developmental forces that
mould a neural circuit responsible for complex vertebrate behavior'
(Arnold 1992). Although conceptually discrete, mechanisms for the
development, production, and perception of song are difficult to separate
in practice. Neural control of song in the Zebra Finch is an are a of
vigorous research activity and only the barest outline of the main
findings wiU be presented here. Fortunately, F. Nottebohm (1993), a
leading researcher in the neurobiology of song learning for the past 25
years, has recently written a masterly review of the important develop-
ments. He compares the neural mechanisms of song Jearning in the
Zebra Finch, an 'age-limited'/'experience-limited' learner, with that of
the canary, a classical 'open-ended' ¡earner, highlighting similarities and
differences and summarising the main questions for future research.
When Price (1979) deafened Zebra Finches late in the sensitive phase
they failed to develop normal songo This is believed to have arisen
beca use they needed to be able to hear what version of song they were
actually producing in order to compare it with the memory of the tutor's
song they previously heard and were trying to copy. The unlearned
elements of song plus the components learned from the tutor during the
10 Vocalisations 239
sensitive phase are believed to combine to form a type of neural repre-
sentation, or template, which the male attempts to copy.
Song production and neural control
Vocalisations in birds originate from the syrinx, a muscular organ
located at the junction of the two bronchi. Each half of the syrinx has a
membrane that vibrates under muscular tension when air is expired from
the lungs so producing sound (Greenewalt 1968). The pattern of muscu-
lar tension on the membranes is caused by nerve impulses which travel
from the brain stem along paired tracheosyringeal nerves to each side of
the syrinx. The production of these nerve impulses is controlled by the
song control circuitry in the forebrain. In addition, the circuitry controls
the respiratory patterns that produce the timing of the songo The struc-
ture and function of this circuit have been under investigation for nearly
three decades and an outline of its components (nudei, or dusters of
neurones) and their connections via nerve projections, as it is currently
understood, is shown in Figure 10.12.
Basically, there are two circuits-the HVC-RA-nXXts circuit controls
the production of song, and the 'recursive loop' is important for learning
of songo Control of the syrinx in the adult emanates from the HVC
which is the pattern generator of the fine detail of the song (Nottebohm
et al. 1990). The HVC and the RA are not connected at hatching, but as
the nestling grows the HVC sends nerve projections to the RA so that the
performance of the first subsong coincides with the initial connection
between the two key control nudei. Subsequent development of song
goes hand in hand with an increase in the size of the HVC and the
strengthening of its link with, and control over, the RA (Nottebohm et
al. 1990). If any one of the control nudei or their projections are experi-
mentally destroyed or damaged, the production circuit ceases aH function
and song is lost. New neurones are not only continually added to the
HVC during maturation, but are also added in the adult Zebra Finch as
well.
The recursive loop also originates at the HVC and ends at the RA but
detours via three control nudei, are a X, DLM, and LMAN. This indirect
circuit is believed to be involved in song learning and perception. Experi-
mental destruction of components of the recursive loop does not stop
production of song in adults, but if destroyed before or during the sensi-
ti ve phase it prevents normal song development (Bottjer et al. 1984).
Where song learning has been experimentally delayed in untutored males,
destruction of the LMAN prevents learning in adulthood (Morrison and
Nottebohm 1993). Electrical recordings in the song control circuits
found that auditory stimuli, especially vocalisations, stimulate neurones
in both production and recursive circuits, with Field L (Figure 10.12)
being a major source of auditory input to the HVC (Katz and Gurney
1981; Williams 1989); these stimuli are transmitted to the brain stem
240 THE ZEB RA FINCH
ANTERIOR POSTERIOR
!s nervB lo
!rachea & Syrinx
Fig. 10. 12 Schematic diagram of the neural system that controls the acquisition
and production of song and learned vocalisations in the Zebra Finch. A sagittal view
of the right side of the brain is shown. Open arrows show the main motor pathway
from the forebrain down to the brain stem and the nerves that go to the syrinx. The
dark arrows show an important alternative pathway, the recursive loop, which is
involved in song learning. Abbreviations of the major control centres, or nuclei, con-
nected by nerve projections (arrows) as follows: HYC, hyperstriatum ventrale, pars
caudale, (or 'higher vocal centre'); RA, robust nucleus of the archistriatum; nXII ts,
tracheosyringeal part of the hypoglossalnucleus; DLM, medial portion of the dorso
lateral thalamic nucleus; LMAN, lateral part of the magnocellular nucleus of the
anterio neostriatum; Y, lateral ventricle. Field L provides auditory input to the HVC
but is not part of the motor pathway. (From Nottebohm et al. 1990.)
Auditory perception
Zebra Pinches do not differ significantly in basic auditory sensitivity from
other species of song birds studied to date (Dooling et al. 1992). Using
positive reinforcement techniques, Hashino and Okanoya (1989) estab-
lished that the range of best hearing for domesticated Zebra Pinches is 1-6
kHz (Figure 10.13). This is the same for most other species of birds that
have been tested so far, and is much inferior to that of humans (Okanoya
and Dooling 1987). Thus, Zebra Pinches are fairly insensitive to sounds
aboye and below this range, but within it they are most sensitive to
sounds around the 4 kHz level which is slightly higher than that for other
songbirds. When Hashino and Okanoya (1989) examined the frequencies
of the Distance Calls they found that most of the sound energy was
between 2 and 5 kHz with a distinct peak at 4 kHz. Coincidence of
frequencies of best hearing and those produced in the vocal utterances is a
common finding among avian species studied to date (Dooling 1982).
The ability to resolve different frequencies against a background
of noise is a more realistic or natural test of perceptual ability, and is
measured by the critical band or signal-to-noise ratio. This increases
monotonically at arate of 3 dB per octave in Zebra Finches, and is the
same as that found for all songbirds and other vertebrate species mea-
~ 80+-~~~++~---+-+~~~
60
40
20
E o
:::J
.........U
al
o..
VJ
1.0 10.0
Frequency in kHz
Fig. 10. 13 Average auditory sensitivity of five domesticated Zebra Finches to pure
tones, and the power spectra of 14 Distance Calls. Sensitivity thresholds were
established by operant reinforcement procedures. (From Hashino and Okanoya
1989.)
246 THE ZEBRA FINCH
sured to date (Okanoya and Dooling 1987). Of course, Zebra Finches
are more sensitive to their own sounds than those of other species when
both are mixed in a sequence together (Okanoya and Dooling 1988,
1990, 1991a,b; Dooling et al. 1992). Furthermore, in one of the few psy-
chophysical demonstrations of sex differences in auditory perception in
higher vertebrates, maJe Zebra Finches were found to discriminate com-
plex heterospecific ca lIs better than females. This may be related to the
fact that males learn to produce Distance Calls and songs and females do
not, and is consistent with the finding of Cynx et al. (1992) described
aboye on sex differences in song discrimination. How specialised percep'
tíon for conspecific calls is acquired is unknown, but the method of
detection is probably a frequency-based filtering system rather than a
'matched-filtering' one involving a signal match with an internal 'temo
plate' (Okanoya and Dooling 1991b).
The minimum detectable threshold for temporal acuity of Zebra
Finches is 2.5 ms for a broad spectrum noise. Within the band of best
hearing, resolution is lowest (3.0 ms) on frequencies centred on 5 kHz
and highest (7 ms) on those centred on 1 kHz (Okanoya and Dooling
1990). As with most species there is a trade-off in resolving power
between frequency and time.
Timbre is a source of much between-individual variation in Distance
CaUs. In addition, Williams et al. (1989) established that the timbre oí
specific elements in the song is under active control of the vocal organ
and accounts for much variation in song structure within and between
individuals. Zebra Finches can perceive differences in timbre and be
trained to discriminate between song elements that vary by as little as 6
dB in just one harmonic out of nine (Cynx et al. 1990). They are much
better at the task when the element is embedded in its normal place in
the song phrase. Inter-hemispheric perceptual differences of Zebra Finch
song have also been detected (Nottebohm et al. 1990).
In summary, temporal and frequency thresholds of Zebra Finches are
about the same as that of other species of songbirds so far studied. How-
ever, their auditory perception is tuned to variations in timbre, especially
those elements given in the song, which is appropriate for a species with
a rich production of harmonics in its vocalisations.
Summary
In addition to the song, Australian Zebra Finches have 12 distinct vocalis·-
ations, of which the Distance Call is the most important. It is the only
long-distance vocal signal emitted by the species. It differs between the
subspecies and sexes, and shows strong geographical and individual vari-
ation. Domesticated male Zebra Finches give abnormal Distance Calls
due to learning errors. The Distance Call in males is learned from the
father early in life, but that of females is not learned. Only four vocalisa-
10 Vocalisations 247
tions of the Lesser Sundas Zebra Finches have been described, but fur-
ther work is needed.
A song contains repeatable phrases, each of which consists of a mean
of six elements or notes in the Australian subspecies and 12 in the Lesser
Sundas subspecies. Songs of wild birds differ geographically, and differ
quantitatively from those of domesticated Zebra Finches. Elements are
sung in a fixed sequence in the phrase and fall into two types: those that
resemble calls and those that do noto Song is believed to have evolved
from a sequence of calls given at take-off. Non-call-like elements are
thought to have originated from the Distance Call. Two types of song
are performed: Directed Song, a pre-coital signal directed at the female,
and Undirected Song which is not directed at any individual, but may
have a number of possible functions, including sexual advertisement and
mate-guarding.
Subsong begins soon after fledging and the plastic song crystallises into
the full adult song around 90 days of age. Laboratory studies of domesti-
cated Zebra Pinches show that they learn most details of their songs in
the first two months of life, but they do not need to learn the
macrostructural components. They learn from the father in the first
month of life, mostly after fledging, but will copy song from other males
during the main sensitive phase (35-65 days of age) depending on the
amount of social interaction they have with their father and with other
males. A range of developmental strategies exist which appear to bias
learning from the father. Nevertheless, only about two thirds of wild
males have a similar song to that of their fathers. In sorne wild families,
song traditions can be detected over three generations.
Song production is controlled by neural circuits located in the fore-
brain. During the first month of life, oestrogens cause the brains of males
to develop song circuits; none develop in females, but if they are treated
with oestrogen, they too develop circuits, and can learn to sing the
father's song and give his Distance Calls. The circuit on the right side of
the brain is more important than the circuit on the left side, and is the
opposite to that found in other species of birds investigated so far. For
song to retain its structural integrity, the bird must be able to hear itself
sing, even when adulto This indicates that song remains plastic after it
has crystallised into the adult formo Song control circuits are not only
involved in the learning and production of song but may have a signifi-
cant role in perception of songo Males and females perceive song differ-
ently and use different parts of the brain. Male perception may depend
on levels of circulating androgens and vary seasonally.
Auditory perception in Zebra Finches is similar to that of other small
birds. Best hearing is between 1-4 kHz, and they are most sensitive to
frequencies around 4 kHz; sensitivity coincides with the frequencies of
the sounds emitted. The minimum detectable threshold for temporal
acuity is 2.5 ms.
11 Sexual imprinting and mate choice
Imprinting
In the mid-1960s Immelmann began to investigate puzzling accounts
from aviculturalists who reported that when they fostered out eggs of
rare estrildines to pairs of Bengalese Finches for rearing, the young, on
reaching adulthood, often formed breeding pairs with Bengalese Finches.
Consequently, in a long series of pioneering experiments, Immelmann
(1969, 1972a) systematicalIy crass-fostered eggs of Zebra Finches
belonging to both domesticated and wild stocks to pairs of Bengalese
Finches that raised the young to independence. y oung were then held
individual1y until maturity whereupon sexual preferences were tested in
'free choice experiments'. Males, which were given a choice of a female
Zebra Finch or female Bengalese Finch at each end of a three-compart-
ment cage (the 'double-choice test'), overwhelmingly directed their
courtship towards the latter. Holding conditions after independence had
no effect on preference; they aH preferred females of the foster-parent
species, even if they could hear and see their own species. The number of
siblings in a braod had no effect on the direction of imprinting, so
Immelmann concluded that young imprint on the individuals that fed
them, namely their parents, not on their siblings.
The sexual preference was highly stable. It persisted in many retests
even after imprinted males had been forced to pair and breed with Zebra
Finch females for periods of up to four years and during which many
broods were successfully reared. Yet, when removed fram their partner
and given a choice of strange females in the double-choice cage they prc-
ferred the Bengalese Finch. Immelmann (1972a) concluded that imprint·
ing was 'absolutely irreversible' since it appeared to last for life no
matter what experiences the male subsequently hado
Clearly, these sexual preferences for the foster parent species began
before independence, and by switching nestlings at different ages fram
their natural parents to the foster-parents, Immelmann determined that
the onset of sexual imprinting occurred between 15 and 20 days post-
hatch. Similarly, by switching independent young at different ages from
the foster-parents back to Zebra Finches he found that it was possible ro
11 Sexual imprinting and mate choice 249
reverse the sexual imprinting on the foster-parent species before 40 days
post-hatch, but not afterwards. He concluded that sexual imprinting
ends before 40 days post-hatch. Immelmann also believed that there was
an own-species bias since reversal of sexual imprinting could only be
made from Zebra Finches to Bengalese Finches if it occurred before day
25 whereas it could occur up to day 40 from Bengalese Finches to Zebra
Finches. Furthermore, if nestlings were foster-reared by a mixed-pair,
one a Bengalese Finch and one a Zebra Finch, 17/21 males became sexu-
ally imprinted on Zebra Finches. However, preferences he re were not as
clear cut and showed signs of 'double-imprinting' (see below).
In contrast to males, female sexual preferences were more difficult to
determine, not only because of their more passive and subtle courtship
behaviour, but their reaction to the stimulus males in the preference tests
were qualitatively different from those displayed by their male counter-
parts (Chapter 9). Nevertheless, in the double-choice test the male of the
foster-parent species received the first courtship greeting, although subse-
quently the Zebra Finch male received sorne greetings if he courted vig-
orously. Females also spent most of the time perched near the Bengalese
Finch. From this small sample (n = 5) Immelmann tentatively concluded
that females also sexually imprint, but have a stronger own-species bias
than males.
Immelmann's inspirational series of experiments encouraged other
researchers, principally J. Kruijt and C. ten Cate at Groningen Univer-
sity, and N.S. Clayton at Bielefeld University to investigate the problem
of sexual imprinting in more detail, especially the causal mechanisms
that lead to the formation of the preference and how this subsequently
affects mate choice. Cross-fostering was again the principal method of
exposing young to the rearing species and while the double-choice test
was not always used it was found to be a reliable method for testing
sexual preference (ten Cate et al. 1989).
Is there an own-species bias?
In follow-up experiments using larger samples, Sonnemann and Sjolan-
der (1977) showed that female Zebra Finches fostered by Zebra Finch
pairs had a very strong sexual bias towards Zebra Finch males, whereas
those fostered by Bengalese Finch pairs also showed sexual preferences,
but these were fairly evenly divided between the two species. These dif-
ferences were interpreted as evidence for an own-species bias and an
hypothesis for an unlearned, or innate preference was proposed where
the effect was stronger in females than males (Immelmann 1972a,b).
Sonnemann and Sjolander (1977) also thought that their results might
reflect the different processes the sexes use to select a partner.
Ten Cate considered the possibility that sources of bias other than a
species effect might also exist in the imprinting and testing processes and
began a series of subtle experiments in which he tried to tease apart the
250 THE ZEBRA FINCH
produce fertile hybrids, whose offspring are also fertile (Clayton 1990d).
How the subspecies discrimina te between one another and why the capa-
bility should exist in the first place, since the nearest populations of the
two subspecies are currently separated by at least 600 km of ocean, are
questions Clayton investigated in a series of complex experiments
(Clayton 1990a,b,c) at Bie1efeld University and whose results are well
summarised in Clayton (1990d).
Both subspecies are morphologically and behavioural1y distinct (Chap-
ters 3 and 10). Playback experiments showed that females preferred
songs of their own subspecies (Clayton and Pro ve 1989) and mate-choice
experiments demonstrated that females preferred males with the carree!
breast band pattern (Clayton 1990a). Thus, at pair formatíon, females
could use song, size and plumage to discriminate among males of the
two subspecies. Clayton (1990b) set out to determine the re1ative impor
tance of these attributes by means of cross-fostering experiments.
When females that had been fostered to consubspecific or hetero-
subspecific pairs to day 35 were subsequently implanted with oestradiol
at adulthood and tested for song preference in playback experiments
they preferred the song of the foster-father's subspecies and did not dis-
criminate between males that were foster-reared or not (Clayton 1990b).
Therefore, females must have learned the macrostructural features of
their foster-father's song befare day 35. This shows the importance of
paternal influence on the learning of the female song preference and is
generalised to songs of other males of the same subspecies at adulthood.
Additional experiments showed that females also learned the songs of
male siblings or aviary mates and this too influenced their song prefer-
ences in adulthood.
In multiple mate-choice tests, narmally raised females of both sub
species preferred normally raised males of their own subspecies over
those that had been cross-fostered. Females must have discriminated
between them using either behavioural eues, or a eombination of visual
and vocal cues (Clayton 1990d). How this preferencc deve10ped was
subsequently examined in eross-fostered and normally reared birds
where the sexual preference was tested in the multiple choice apparatus
and subsequently eonfirmcd in pair-formation triaIs conducted in
aviaries (Clayton 1990a). Preference of the normal1y reared group was
stronger than the foster-reared group so that during pair formatíon pref-
erences of the former prevailed over those of the latter, especially in
females, sinee they have a more decisive role in pair formation than
males (Chapter 9). The important finding is that the behavioural differ-
ences in subspecies sexual preference have arisen from sexual imprinting,
and this is probably the factor that maintains sexual isolation between
Australian and Lesser Sundas Zebra Finches in captivity. Plumage and
macrostructural song cues are both learned for subspecies recognition,
but neither are altered by rearing experience.
11 Sexual imprinting and mate choice 257
Clayton (1990d) considered that the cues used for assortative mating
of subspecies might also be available for mate choice within species. For
example, macrostructural differences in element number are used to dis-
tinguish subspecies but the same cues are also used to discriminate
among singers of the same subspecies in playback experiments (Clayton
and Prove 1989). Thus, the same cue is used for both sexual imprinting
and sexual selection. Bill colour and breast band size may also fall into
this category.
Mate choice
At pair formation, Zebra Pinches usually have a choice of unmated indi-
viduals with whom they can pair and it is critical that they make the cor-
rect choice, especially since the pair bond lasts for life, and it is rare for
individuals in the wild to outlive more than three or four partners (Chap-
ter 6). This means choosing and pairing with the highest 'quality' indi-
vidual available in order to maximise fitness. Presumably, phenotypic
indicators of gene tic quality and reproductive and rearing competence
should be the cues used to attract partners. Optimising choice of a part-
ner should be a concern for males as well as females, since males make
an almost equal contribution to the reproductive effort of the pair, and,
indeed, this is the case (Chapter 9). Nonetheless, Zebra Finches are sexu-
ally dimorphic indicating that sorne form of sexual selection is operating.
For this reason they have been chosen as subjects by evolutionary biolo-
gists, notably by N.T. Burley and her students in the United States, who
are interested in mate choice and sexual selection in monogamous
speCles.
Choice of mates and copulating partners
Sex ratios in flocks and breeding colonies fluctuate from month to month
in response to changing patterns of mortality and immigration (Chapter
8) so that the extent of choice varies significantly from day to day and
may even be affected by seasonal factors. Occasionally, through lack of
available options, sorne individuals may be forced to pair with partners
of lesser quality. Under these circumstances it is adaptive to seek matings
outside the pair bond with individuals that are superior to the current
partner and this presents birds with another choice, namely, which mem-
ber of the opposite sex should be chosen for these extra-pair copulations.
Opportunities for extra-pair copulations will be high for males and fairly
continuous throughout the long breeding season, except for the fertile
period of their own partners when they are constrained by supply of
sperm and mate-guarding duties (Chapter 9). Benefits of extra-pair copu-
lation are obvious for males in that they can increase their reproductive
success without any additional parental investment. Females, in contrast,
have fewer opportunities for extra-pair matings given the close guarding
258 THE ZEBRA FINCH
by the partner and the fact only about one to three fertile periods occur
each breeding season. A number of benefits have been hypothesised for
female extra-pair copulations, including the most likely ones of 'good
genes', and 'genetic diversity' (reviewed by Birkhead and M011er (1992)),
but the pattern of occurrence of extra-pair copulations from the Danaher
colony do not support either of these (Birkhead et al. 1990). The relative
importance of the two types of mate choice should also be kept in
perspective: DNA finger printing at the Danaher colony found that only
12% of females had extra-paír offspring and these constituted onIy
2.4% of all offspring sampled (Chapter 9). Thus, the choice of a mate is
a critical decision for both male and female Zebra Finches.
Laboratory experiments on domesticated Australian Zebra Finches
have investigated both types of heterosexual choices: partners for the
pair bond and partners for extra-pair copulations. Choice of paír bond
partners has been investigated in multiple-choice arenas that allow two
way interaction between the choosing individual and those offered for
choice. Preferences are normally measured by timing how long the
choosing bird spends in front of the cages of the various individuals on
offer and subsequent experiments have shown that this measure of
attractivencss is a true indicatíon of the pairing preference (Burley et al.
1982; Clayton 1990a,b); similarly, scoring of Directed Song is another
means of determining pairing prefcrence (ten Cate 1985b). There are
three important questions: is there agreement on which individuals are
the most preferred, and if so, what cues make them attractive, and are
these cues reliable indicators of quality? So far researchers have concen-
trated on female preferences for male Zebra Finches.
Female Zebra Finches require males of a minimum standard of quality
but their preference for a mate is relative to recent experience (Collins
1995). That is, if a female had recently experienced a high quality male
she will not accept a lesser quality one shortly afterwards.
Females prefer males with deep red bills
Females prefer those males with the reddest, brightest bill colours, and
males that had the bill painted to a 'super red' shade were preferred over
males that were painted with deepest naturally occurring shades (BurIey
and Coopersmith 1987; but see Sullivan (1994) and Weisman et al.
(1994) where orange-billed males were preferred). Hence, for females,
the redder the bill the better, whereas males prefer females with the stan-
dard orange-red colour and avoid females whose bills are too red 01' too
yellow. In a similar series of preference tests, in which bill colours of
males were not artificially altered, Houtman (1992) likewise found a
positive relationship between bill redness and attractiveness. Rate of
singing was also positively related to attractiveness, and could compen-
sate for an unattractive bill colour (Collins et al. 1994).
Is bill colour an indicator of male quality? Colour of the bill in wild
11 Sexual imprinting and mate choice 259
and domesticated Zebra Finches varies among individuals but is corisis-
tent within them, although it changes during the course of reproduction
(Burley et al. 1992). Bill colours of domesticated and wild Zebra Finches
have a similar range and distribution of shades; they are redder and
darker (higher colour scores) in males and more orange and paler (lower
scores) in females, but there is some overlap between them. Bill colour
scores of wild birds decline over the breeding season in populations from
central Australia and northern Victoria (Burley et al. 1992; Zann
1994b). Among laboratory birds, significant declines in bill colour and
weight occur over a five-week breeding cycle in both sexes, and colours
quickly recover if breeding is prevented. Other factors also affect bill
colour. For example, at the end of a drought in central Australia all birds
had pale bills, but these increased in colour when birds were held in brief
captivity on an enriched dieto Given the aboye effects, Burley et al.
(1992) concluded that bill colour is an indicator of quality but is affected
by physical condition which is lowered by reproductive effort and harsh
living conditions. However, contrary to expectation, bill colour in wild
females was positively correlated with ectoparasite loads, but experi-
ments are needed to determine how parasite load affects the physiology,
bill colour and mate choice before this finding can be fully interpreted
(Burley et al. 1991). Despite the dynamic quality of bill colour, the opti-
mally preferred colour for members of the opposite sex is attained just
before, and during egg-laying, that is, during the female's fertile periodo
This is the time when both partners need to be at their most attractive;
the male needs to keep his mate from seeking extra-pair copulations, but
if she does, she may need to attract the best quality maleo
In an aviary study, Burley and Price (1991) found that females invited,
or tolerated, extra-pair copulations from only a subset of those males
available, namely those with redder bills and larger breast bands than
possessed by the female's own mate. Bill colour is strongly heritable in
domesticated Zebra Finches, but evolution is only proceeding slowly
towards optimal sex-specific colours beca use it is constrained by low
dominant genetic variance and opposing selection in males and females
(Price and Burley 1993).
Females prefer males with high song rates
In multiple choice tests, Houtman (1992) found that females preferred
some males over others and there was agreement on which males were
most attractive. These were males with high rates of directed singing.
Later these same females invited extra-pair copulations from males that
had higher rates of singing than their own partner, while those with
lower rates were ignored or deterred. High song rate was not a conse-
quence of female preference (Collins 1994). In follow-up cross-fostering
studies Houtman (1992) found significant heritability in singing rates
between father and son and between brothers. Consequently females that
260 THE ZEBRA FINCH
copulated with frequently singing/courting males could expect to have
sons that also sang frequently and were attractive to females. Whether it
is the song component of the courtship display and/or the dance compo--
nent that makes males attractive is unclear. Playback experiments by
Clayton and Prove (1989) showed that females prefer long complex
song-phrases. However, in these experiments song rate was controlled.
Singing rate, like bill colour, is high in males in good condition. Hout-
man (1992) found singing rate was positivcly corrclated with the amount
of fat stored around the clavicle, which is one indication of physical con-
dition. Another corrclation was found between singing rate and the
mean weight of the singer's offspring at independence. Presumably, only
competent fathers in good condition can produce heavy offspring to
independence.
Sheridan (1985) attempted to determine which component of
courtship females found the most attractive, the song or the dance.
Despite efforts to court and sing, castrated and muted males were not as
attractive as intact males which courted more frequently and gave the
full display.
In a semi-natural aviary study, Ratcliffe and Boag (1987) could find no
corrclation between song rates and male mating success; instead, aggres-·
sion (supplantings) was strongly correlated with success. Their experi-
ments were designed to heighten aggression and competitiveness among
males and this may have obscured any effect of singing.
Effects of hormones on male attractiveness
Rate of singing and bill colour were positivcly corrclated in Houtman's
(1992) study. This is not surprising beca use both depend on levels of
testosterone in the blood. In laboratory and wild Zebra Finches rates of
Directed and Undirected Song were positively correlated with testos--
terone levels (Chapter 10), theoretically enabling females to assess the
underlying physiological sta te of males they encounter. Predictably,
when males are castrated they lower their rates of singing and courtship;
if subsequently implanted with testosterone, the effects are reversed
(Arnold 1975a; Harding et al. 1983). Similarly, after castration the bill
colour changes from red to the orange colour typical of females, but
testosterone implants quickly restore the red colour typical of males
(Cynx and Nottebohm 1992b). Sheridan (1985) enhanced the attractive--
ness of males by implanting androstenedione (an androgen), but could
find no increases in singing and dancing performance that could be serve
as cues for the females; unfortunately, he did not examine changes in bill
colour, which may be more sensitive to hormones than sexual behaviour
and could have been the cue used by females in this experimento In both
sexes there is a need to examine in detail the nature of interactions
among hormones, physical condition, bill colour and courtship
behaviour.
11 Sexual imprinting and mate choice 261
of nestling Begging Calls differed between the sexes. Nestlings of first and
second generatíon offspring of wild-caught Zebra Finches also have sex
differences in Begging Calls; specifically females have a greater range of
caIl variants than males, but no ca lIs are unique to one sex (Roper 1993)0
Brood reduction does occur in wild Zebra Finches (Chapter 9), but
there is no evidence for or against facultative manipulatíon of numbers
of nestlings in broods. However, if resources are limiting and it is neces-
sary to reduce the size of the brood, selection should favour a mechanism
to bias the sex ratio towards that sex that would achíeve the híghest fit-·
ness. Under most conditions the sex of offspring likely to inherit the
attractive characteristics of the most attractive parent should be
favoured, although in locations whcre there is a local imbalance of adults
of one sex (see Burley et al. 1989) fitness would be higher by favouring
the nestlings of the sex in most demando
Until a cheap molecular technique is available to sex recently hatched
Zebra Finch nestlings it is difficult to envísage field investigarions of sex
ratio bias comparable to those performed by Burley. High rates of nest
predarion and fledgling mortality would make gathering of such data
arduous at best.
Summary
Zebra Finches are monogamous and form lifelong bonds in which both
sexes participate in all aspects of parental careo Consequently, choice of
the best qua lit y partner is of major importance for the fitness of males
and females. Members of the opposite sex are also chosen for participa-
tíon in extra-pair copulations. Females prefer males with deep-red bilis
for both paír formation and for extra-paír copulations. Redness of the
bill ís inherited and indicates physical condition, but is subject to fading
as a consequence of reproductive effort and harsh environmental condi-
tions. Females are also attracted to those males that show courtshíp
vigour and sing frequently; these males are preferred for extra-pair copu-
lations. A high song rate is heritable and is also correlated with good
physical condition. Both red bills and high song rate are also correlated
with high levels of testosterone so that both attributes may indicate good
physical condition, strong courtship vigour and high intra-male competi-
tíve ability.
Despite visual and auditory cues for kin recognition, Zebra Pinches
have no biases for, or against, paíring with relatives, and a brother-sister
mating has been found in free-living birds. High mobility and dispersal
makes sib-sib pairing infrequent. Inbreeding depressíon in domesticated
Zebra Pinches has not been c1early established.
Wild-caught and domesticated Zebra Finch females are attracted to
males wearing red bands and avoid males wearing light-green bands;
males are attracted to females wearing black bands and avoid those with
11 Sexual imprinting and mate choice 269
light-blue bands. Colour preferences may be based on biases towards
their own species-typical colours and against species-typical colours
found in closely related species. Females prefer males that have bands
placed in bilaterally symmetrical arrangements. It is hypothesised that
males of superior quality have their secondary sexual characteristics
arranged in a more symmetric fashion than do inferior individuals and
that band preferences are an expression of this. In domesticated Zebra
Finches, coloured leg bands affect mate selection, extra-pair copulations,
parental investment, reproductive success, sex ratio s of offspring and
longevity. Experiments on wild Zebra Pinches did not detect colour band
effects on mortality or physical condition, perhaps because breeding sea-
sons were not as long as those possible in captivity. However, in wild
pairs where males have red bands, there was a significant increase in the
reproductive success (eggs, clutches). In order to prevent behavioural
artefacts in field and laboratory studies, red bands are not recommended
for male Zebra Finches.
12 Life history and adaptations
'The large fluctuations of the Zebra Finch population at Jindi Jindi resembles
that of the grasshoppers and suggests that both breed prolifically when
conditions are good; most die but a few survive over the dry years. Autumn
breeding is presumably an advantage in a species that can breed at three months
of age because~ when numbers are very low~ the survival of even a few will
boost the potential of the breeding population in the spring.'
s. J. J. Po Davies (1986).
The primary aim of the final chapter is to integra te the various anato--
mical, physiological and behavioural attributes of Zebra Finches in
order to understand the principal thrusts of their life histories and
adaptations.
Both subspecies of Zebra Finches must be considered successful in
their respective environments beca use both are re1ative1y numerous,
wide-spread in distribution, and persistent in a rapidly changing environ·
mento Understanding their life histories is the key to understanding why
they are successful. Unfortunate1y, few details of even the most basic
aspects of the fie1d biology of the Lesser Sundas subspecies are known.
Consequently, little can be said about its microevolutionary adaptations
to a highly seasonal environment. Whereas Immelmann (1962a, 1965a)
described the basic breeding ecology of almost all of the Australian
estrildines, little has been published on their demography, with the
exception, of course, of the Zebra Finch. Demography is the key to
understanding life histories (Stearns 1992), but unfortunate1y demo--
graphic data are only available from just a few locations within the con-
tinent-wide range of the Zebra Finch so that the extent of local
adaptations in demographic and other traits in different environments
are only vaguely suggested at this stage.
The demographic traits listed in Table 12.1, measured principally from
the Danaher colony in northern Victoria, show that the Zebra Finch life
history is distinguished by an exceptionally earIy age of maturation, high
leve1s of pre-adult mortality, and multi-broodedness.
Precocial breeding
Age of maturation is a pivotal trait in life history evolution beca use fit·
ness is more sensitive to this trait than any other (Stearns 1992). The
benefits of early maturation are shorter generations and higher survival
12 Life history and adaptations 271
Table 12.1 Demographic traits significant in the life history of Australian Zebra
Finches.
Mortality patterns
Nest predation (66%) is quite high among Zebra Finches breeding at
the Danaher colony in northern Victoria (Table 12.1) and is comparable
to that found among breeding birds at Alice Springs (R. Zann and
N. Burley, unpublished observations). In general, intensity of nest
predatíon of passerines is hígh in Australia and varíatíon in reproductive
effort of the 'old Australian endemic' species of passerines occurs
through number of breeding attempts rather than c1utch size (Rowley
and Russell 1991). High re-nesting rates in Zebra Finches may be a
response to predation levels, and the fact that their eggs are smaller than
expected for estrildines of their mass (Chapter 6) may be an adaptatíon
for this.
Nest predatars in different parts of the range appear to be well
adapted to exploiting Zebra Finches. They not only take a heavy and
constant toll of eggs and nestlings from permanent1y established colonies
that have regular breeding seasons, but in regions where breeding is
unpredictable in time and space predators quickly track and find nesting
colonies and rapidly exploit them.
The next most distinguishing feature of the Zebra Finch life cyele,
after the early age of maturation, is the extremely high rate of juvenile
mortality. The proportion (22%) of fledglings that appear to reaeh
breeding age (day 80) is well below the 42-86% range for a number of
different passerine species summarised by Newton (1989). Risk to juve-
niles is highest between fledging and day 35, the age of nutritional inde-
pendenee (67% losses), but after independenee they grow out of risk so
that losses between independence and day 80, the age of sexual matura-
tion, fall to 27%, whieh is not distinguishable from that of adults (see
Figure 8.2). Losses before day 35 are most1y mortalities sinee dispersal
by young before that age is unlikely; however, estimates of losses from
independenee to sexual maturity confound disappearanees due to disper-·
sal with those due to mortalities. Nevertheless, mortalities before matu··
ratio n must be very high and 'instantaneous juvenile mortality' is
believed to be the majar seleetive force in promoting and maintaining
early maturation (Stearns 1992).
In modelling the relationship between environmental eonditions and
early maturation, Stearns and others assumed that growth rate of juve-
niles was a potential eue for monitoring prevailing demographie condi-·
tions: when growth slowed down there were inereases in juvenile
marta lit y rates and this was eompensated by delayed sexual maturation.
However, early sexual maturation in Zebra Finehes appears be a fixed
trait across its geographie range and, although demographíe data exist
only for the Danaher colony, growth rates were not notieeably affeeted
by environmental eonditions, although there was a delay in the onset of
moult in the second half of the breeding sea son (Chapter 3). Further
12 Life history and adaptations 275
study is needed to determine if the growth rate of potential breeders is a
cue to undertake precocial breeding or noto
Among estrildines, early maturation is the derived trait and later matu-
ration the ancient trait fixed in the lineage, hence early maturation of
Zebra Finches must be an adaptation to unique ecological conditions. In
estrildines other than the Zebra Finch, delayed maturation must reduce
the rate of juvenile mortality sufficient to compensate for the increased
risk of having to survive for a longer period befo re reaching breeding
age. The environment of the Zebra Finch is distinguished by long breed-
ing seasons or episodes and high levels of juvenile mortalities and both
factors would have selected for early sexual maturation. Juvenile mortal-
ities probably arose mainly through extrinsic factors-those not sensitive
to changes in reproductive decisions. Starvation and predation are the
main sources of extrinsic mortalities in Zebra Finches, and the effects of
reproductive costs on ageing constitute the main intrinsic so urce of mor-
talities. I argued in Chapter 7 that reproductive costs in Zebra Finches
are not normally important, except when many successive broods are
raised without interruption, such as in captivity, or during exceptionally
long breeding episodes that may occasionally arise in central Australia;
otherwise, cessation or decline in breeding conditions stops further
breeding attempts and minimises long term costs. After early maturation
evolved, other selection pressures maintained it by reducing the level of
juvenile mortality, principally through high reproductive competence of
young novice breeders.
What prevents Zebra Finches from attaining sexual maturation earlier
than they do already? Energetic constraints imposed by moult are proba-
bly the main factors. Moult into adult body plumage begins around day
35 and is not complete before about day 70. Failure to develop complete
foraging skills may be another factor in preventing the advancement of
breeding age, beca use techniques needed to take rearing food, such as
half-ripe seeds from standing heads of grass, may take sorne time to
develop. Fortunately, the more complex hunting skills needed to take fly-
ing insects are unnecessary.
Drinking
The tip-down rnethod of drinking in the Zebra Finch and seven other
Australian species is believed to be an advantage in arid environrnents
(Chapter 5). Heidweiller and Zweers (1990) postulated that the cornplex
tip-down drinking evolved frorn the tip-up rnethod found in all other
species of estrildines and that its evolution was associated with an eco-
logical shift in habitat frorn rnesic to serniarid and arid environrnents
with an associated switch to a diet of grass seeds. The anatornical and
behavioural adaptations for dehusking grass seeds served as pre-adapta-
tions for tip-down drinking, but why this should have evolved in the
Australian estrildines and not African ones is unknown. Irnrnelrnann and
Irnrnelrnann (1967) believe it rnay be associated with the greater insectiv-
orous diet in the latter, where sorne arid-adapted species (e.g.
Uraeginthus granatina, Pytilia melba) can go for rnonths without drink-
ing, providing insects are plentiful.
Phylogenetic analysis of the rnethod of drinking is sornewhat arnbigu-
ous, but it suggests that tip-down drinking evolved twice in the Poephilini
(Figure 12.1). It is an ancient character that originated early in the evolu-
tion of the tri be sorne time after the ancestor of the Painted Finch sepa-
rated off frorn the cornrnon ancestor of the other rnernbers and therefore
it is the ancestral condition as far as Zebra Finches are concerned.
Nest roosting
Desert environrnents are renowned for low nocturnal ternperatures and
Irnrnelrnann (1962a) suggested that roosting in nests is one rneans of
conserving heat arnong desert inhabitants. In the dry season in tropical
northern Australia, ternperatures also fall sharply at night. Indeed,
Davies (1986) believes that low ternperatures in Australian deserts and
arid regions provide considerable environrnental stress for anirnals.
Estrildines, with their high zone of therrnal neutrality, are particularly
susceptible to cold, and Zebra Finches can rnake significant energy sav-
ings by roosting in nests (Chapter 6). The incidence of nest and perch
roosting in the subfarnily was extracted frorn the literature (Goodwin
1982; Irnrnelrnann et al. 1963, Irnrnelrnann 1965a), and although nest
roosting is also found in sorne estrildines outside the Poephilini, espe-
cialIy in sorne African genera (Spermestes and Amadina), character opti-
rnisation in both the Lonchurini and Estrildini showed that perch
roosting is plesiornorphic for the Poephilini. However, phylogenetic anal-
ysis shows that nest roosting is not a recently evolved trait in the
Poephilini, but arose in an early ancestor shortly after the speciation of
Emblema pictum (Figure 12.1).
There is sorne congruence in the patterns of distribution of the roost-
ing and drinking characters when rnapped onto the phylogeny of the
280 THE ZEBRA FINCH
Habitat NSiM MiS S M M SiM SiM SiM M S SiM M A
Seasonality NS S S S S S S S S S S S A
X T.9. T.b. 5.9. S.b. 5.0. P.p. P.a. P.c. N.m. N.p. E.p.
~
~uJ
Fig. 12.1 Method of drinking and roosting mapped onto a phylogenetic tree of the
Poephilini constructed by Christidis (1987b) using chromosomal and biochemical
characters. Drinking and roosting were polarised using the sister group (X), the
Lonchurini (see Fig. 1.1), as the outgroup. Tip-up drinking and perch roosting are
ancestral traits (plesiomorphies) in the Poephilini, and tip-down drinking and nest
roosting are derived traits (apomorphies). T.g. == Taeniopygia guttata; T.b. =: T.
biehenovii; S.g. == Stagonopleura guttata; S.b. == S. bella; S.o. == S. oeulata; P.p. :"
Poephila personata; P.a. == P. aeutieauda; P.c. == P. cineta; N.t. == Neoehmia tempo-
ralis; N.m. == N. modesta; N.r. := N. rufieauda; N.p. == N. phaeton; E.p. == Emblema
pietum. Habitat: A == arid, M == mesic, S == semiarid. Seasonality: A == aseasonal, S :=
seasonal.
Breeding potential
In addition to precocial breeding, Zebra Finches possess a number of
traits that give them a high breeding potential which is utilised when
12 Life history and adaptations 281
suitable conditions arise. Sorne are unique to the species, whereas others
are found to varying degrees in other estrildines. The ability to breed at
anytime of the year is a consequence of a tonic gonadotropic activity of
the hypothalamo-hypophysial system which maintains the gonads in a
semi-permanently active state (Chapter 7). To date this has not been
demonstrated for any other estrildine. In contrast, other characteristics
of Zebra Finches that facilitate opportunistic breeding may be pre-adap-
tations, in the sense that they may ha ve evolved in an ancestral species.
For example, the ability to moult and breed at the same time is also an
advantage for an opportunistic breeder, but may not have arisen as a
microevolutionary adaptation in the Zebra Finch, since it occurs in at
least six other species of Poephilini that live in semiarid habitats
(Schoepfer 1989; Tidemann and Woinarski 1994). Similarly, the tight,
life-Iong bond is an advantage for an opportunistic breeder since it expe-
dites ovulation and accelerates breeding. However, a life-Iong bond
appears to be a characteristic trait of the subfamily, although only a few,
the Poephila, the Amadina and the Uraeginthus species, have a pair bond
that is as tight year-round as that of the Zebra Finch. Interestingly, all of
these species are arid-adapted and appear to have developed the trait
independently. Finally, slow wing moult may not be a microevolutionary
adaptation of Zebra Finches that facilitates mobility and dispersal since
it occurs in the Double-barred Finch (Schoepfer 1989), and also in
Masked and Long-tailed Finches (Woinarski and Tidemann 1992) that
are not highly mobile. However, the ability to moult continuously
(Chapter 3) has not been found in any other estrildine and may be
another trait that adapts Zebra Pinches to their unique predominantly
aseasonallife cyele.
Sexual dimorphism
Of all estrildines, few species are as strongly sexually dimorphic as the
Zebra Finch. Size dimorphism is extremely rare in estrildines, and only
occurs in the Pin-tailed Parrot-Finch Erythrura prasina, where the tail of
the female is only about half as long as the male's (Goodwin 1982).
Although the sexes of Zebra Finches differ in size, the differences are
only slight statistical ones (Chapter 3), and the species should be consid-
ered sexually monomorphic in the strict sense. Where secondary sexual
characteristics exist in estrildines, there are differences in colouration in
plumage, bill and soft parts; that is, they are sexually dichromatic.
Table 12.3 shows that Zebra Finches are the most sexually dichro-
matic of estrildines, with four to five male-specific plumage markers
having evolved in the Australian subspecies (ear, flank, chest, throat, and
possibly abdomen), and three in the Lesser Sundas (ear, flank, chest).
The location of these sex-specific markers makes them highly cons-
picuous when viewed from the front or side, yet they are invisible from
282 THE ZEBRA FINCH
Table 12.3 Levels of sexual dichromatism in estrildines based on illustrations and
species descriptions of mature adults taken from the !iterature (mainly Immelmann
et al. 1965a; Goodwin 1982)
._---------------_._--
Sexual chromatism Description Examples
for Zebra Finches and their patterns of parental investment are consis-
tent with this view. However, evidence for intersexual selection comes
from both captivity and field studies where females have been shown to
discriminate among males for the purposes of pair formation and extra-
pair copulations (Chapters 9 and 11).
What behavioural and/or ecological factors selected for and maintain
sexual dichromatism in Zebra Finches and a handful of other estrildines,
and what has prevented other species of estrildines evolving strong sex-
ual dichromatism? Phylogenetic analysis suggests that strong sexual
dimorphism evolved independently three times among the Poephilini: (a)
in the Zebra Finch, (b) in the ancestor of the Neochmia clade
(monochromatism evolved back in N. temporalis), and (c) in the Painted
Finch (Figure 12.2). With the exception of E. prasina, estrildines with
strong or extreme sexual dichromatism live in arid environments, but
only the Zebra Finch and the Painted Finch have largely unpredictable
breeding seasons. Possibly, the ancestor of the Neochmia was also an
inhabitant of arid environments.
Seasonality AlS S S S S S S S S S S S A
X T.g. T.b. 5.9. S.b. 5.0. P.p. P.a. P.c. N.m. N.r N.p. E.p.
Fig. 12.2 Degree of sexual dichromatism mapped onto a phylogenetic tree of the
Poephilini constructed by Christidis (1987b) using chromosomal and biochemical
characters. Sexual dichromatism was polarised using the sister group (X), the
LOl1churini (see Fig. 1.1), as the outgroup. O-sexual monochromatism, 1-slight
dichromatism, 2-moderate dichromatism, 3-strong dichromatism, 4-extreme
dichromatism (see Table 12.3). Abbreviatiol1S of specics names and habitats as in
Fig. 12.1.
12 Life history and adaptations 285
Burley (1981b) predicted that sexual indistinguishability is an advan-
tage in monogamous, social species, especially those that flock in the
breeding season, or breed colonially. A literature and museum survey of
150 species supported her prediction, but there were three notable excep-
tions, the Flock Pigeon, the Budgerigar and the Zebra Finch. The need
for a rapid breeding response was thought responsible for selection for
strong dichromatism in these arid-adapted species. Burley argued that it
was an advantage for individuals that bred in colonies to conceal their
sex in order to reduce the level of sexual interference competition and so
enhance fitness. Sexual interference occurs in Zebra Finch colonies but
seems limited to only a few days in the fertile period: copulating males
may be knocked off by others and rates of extra-pair courtship and
forced copulations suggest that male interference of females can be quite
high (Chapter 9). It is surprising that Zebra Finches have not adopted
nest copulation where it would be free from interference from other
colony members since it has evolved a number of times in at least a
dozen species of estrildines from all three tribes (Immelmann and Immel-
mann 1967). Interestingly, most species of nest copulators are highly
social and sexually dichromatic ones. Sexual indistinguishability does not
necessarily prevent sexual interference and promote social harmony. For
example, white Zebra Finches are much more socially disruptive than
wild-type Zebra Finches, although this arises from incompatibilities
between external appearance and behavioural responses. Similarly, field
observations of Poephila cincta, a species where the sexes are barely dis-
tinguishable, shows that defence of the mated partner from sexual com-
petitors is high (Zann 1977) and suggests that where permanent colonies
are established, sex is difficult to conceal even when the sexes appear
indistinguishable to humans.
Conservation
Human activities in Australia and Indonesia have increased the distribu-
tion and abundance of the species, although protection from trapping
in the latter country may be needed. The extent of available habitat
has in crea sed through clearing of forests and woodlands and changes in
land use and cultivation have increased the availability of food plants
(Chapter 2).
Since European settlement of arid Australia, thousands of artesian bores
have been drilled for the pastoral industry and these have inadvertently
provided drinking water for dispersing Zebra Finches and allowed them
to establish semi-permanent populations in areas not previously occupied
(Davies 1977a). Currently however, there is a program to conserve the
artesian supplies by systematically capping many degraded and wasteful
bores not used by stock; this will exclude Zebra Finches and other native
wildlife that require drinking water from sorne localised areas.
286 THE ZEBRA PINCH
In southwestern and southeastern Australia where clearing of wood-
lands and maUee for cultivation has extended Zebra Finch habitat, lack
of suitable indigenous nesting bushes is limiting in sorne regions espe-
cially where thorny exotic species, such as boxthorn are subjected to
eradication.
Grazing by exotic herbivores in the vast mulga shrublands of arid
Australia has degraded much of the chenopod understory. Water
previously used by chenopods has become available for growth of annual
grasses which now predominate the understory (Davies 1977a). Accord-
ingly, therc has been a great increase in the distribution and abundance
of seed food for Zebra Finches and other granivorous species (Reid and
Fleming 1992). Nonetheless, overgrazing by catde and rabbits of these
grasses especially those growing on floodplains and riverine environ-
ments has led to serious degradation of vast areas of rangelands. In
sorne cases highly valuable drought refugia, centred on rare 'islands' of
nutrients and moisture, have been irreversibly destroyed by overgrazing,
and in other cases this has allowed woody weeds to take over and pre-
vent regrowth of grasses. Thus the carrying capacity for grazers and
granivores alike has been permanently reduced in sorne places. Heavy,
sustained use of rangelands by catde, rabbits and kangaroos is an impor-
tant conservation problem in Australia and hopes of restoration rest
mainly on the elimination of rabbits and lower stocking rates of catde,
especially during drought.
Thankfully, however, the conservatíon of Zebra Finches is not a
current concern, nor likely to be one in the near future for either sub-
speCles.
Summary
Many of the traits that enable Zebra Finches to thrive in an unpre-
dictable arid environment have arisen in distant ancestors and are
lineage-·specific adaptations that have not originated as a consequence of
microevolutionary selection on Zebra Finches or their immediate ances-
torso Many physiological (high body temperature, low metabolic rate,
high thermal conductance), morphological (small body size) and life his-
tory (graminivory, life long paír bond, life span) traits characteristic of
the Estrildini are adaptive to arid environments. This suggests that the
proto-estrildine probably evolved in an arid or semiarid environment. In
addition to these pre-adaptations other arid-adapted traits probably
evolved in the ancestor of the Poephilini (nest roosting, tip-down
drinking, slow moult, breeding while moulting) and were maintained in
descendent species during the clima tic changes in Pliocene-Pleistocene
epochs. Finally, there are those unique traits that arose in Zebra Finches
as microevolutionary adaptations to an arid, unpredictable environment.
Early sexual maturation, precocial breeding, constant gonadal activity,
12 Life history and adaptations 287
opportumstlC breeding, ability to withstand extremes of dehydration,
high mobility and high dispersal are the distinguishing adaptations of
Australian Zebra Finches. These traits enable survivors of perennial
droughts to find and exploit any favourable conditions, and by means of
a prolific capacity for breeding, populations can be quickly restored
befo re the next devastating drought. Field work is needed on the Lesser
Sundas subspecies of Zebra Finch in order to determine how these adap-
tations to aridity have been modified in the monsoonal seasonal enviro n-
ment that prevails throughout its range.
Appendix 1
POACEAE
Amphibromus neesi SE4 P Sp-A
Aristida contorta WA2, e 7, EA 3 AlP Sp,A
Aristida armata EA3 P A
Aristida inaequiglumis e7, 3K5 P ?
Aristida muricata EA P A
Astrebla lappacea EA 3 P
Astrebla pectinata EA3 P
Arundinella nepalensis K1
Avena sp.b SE4 P Sp
Botrichloa macra SE6 P SulA
Brachyachne convergens K1,5 A Su
Brachiaria gilesii e77 A A
Brachiaria miliiformis e7 AlP SulA
Cenchrus ciliaris b e4 P SulA
ehloris truncata SE AlP Su
ehloris viragata b K1 P SulA
Cymbopogon obtectus e77 4 P Sp/A
Cynodon dactylon e 3SE P SulA
Dactyloctenium radulans EA A Sp/Su
Danthonia caespitosa SE4 P Sp
Danthonia sp. SE6 P
Dichanthium sericeum e 77, EA 3 P Su
Dichanthium tenuiculum e7 ? ?
Digitaria brownii e P Sp-A
Digitaria sanguinalis b SE4 A SulA
Echinochloa colonum b e2,4 K5 A SulA
Echinochloa crus-gallib SE A SulA
Enneapogon avenaceous EA3 A SulA
Enneapogon cylindricus e 7, EA3 AlP SulA
Enneapogon oblongus e 71 P ?
Enneapogon pallidis K P SulA
Enneapogon polyphyllus e57, EA3 AlP Su/W
Enneapogon glaber K
Eragrostis australis WA2 P Sp
290 Appendix 2
----------
Species Location and Growth habite Floweringd
sourcea
--------------------
7
Eragrostis eripoda e , P A
Eriachne aristidea EA 3 AlP Sp-A
Eriachne sp. WA 2
Erichloa pseudo-acrotricha SE6 P SulA
Iseilema membranaceum 2 2 3
e , K , EA A SulA
Neurachne michelliana 7
e P W
5
Oryza rufipogon K
Panicum decompositum EA 3 P SulA
Panicum prolutum SE4 P Sp
Panicum zymbirformae K2
7 2
Paspalidium sp. e , K , AlP Su-W
6
Paspalidium criniforme SE P SulA
Paspalidium dilatatum b SE4 ,6 p Sp-Au
Phalaris sp. b SE4
Poa annua b SW2 , SE4 AlP W/Sp
Sorgum stipoideum K5 A Su
2
Sporobolus australasicus e , K2 A
6
Stipa densiflora SE P Sp
Stipa seta cea SE6 P W/Sp
Themeda australis e7 p Sp/Su
Themeda triandra K5
Thyridolepsis michelliana el p Sp-A
Tragus australianus el A Su-Au
Trirapsis mollis EA 3 P Sp-Au
Triodia basedowii el p Sp-W
5
Triodia wiseana K
7
Tripogon loliiformis e A S/W
CHENOPODIACEAE
Maireana scleroptera (} P
EUPHORBIACEAE
Euphorbia drummondii el A
POLYGONACAE
Polygonum aviculare SE4
---------------------"-------".-
aLocations: C, Central Australia; K, Kimberley and western Northern Territory; EA,
eastern arid zone, WA, western arid zone; SE, southeastern Australia; SW, south-
western Australia; Sources: 1 Immelmann 1962a; 2, Davies 1977a; 3, Morton and
Davies 1983; 4, Zann and Straw 1984a; 5, Tidemann 1987; 6, Schopfer 1989; 7,
Zann et al. 1995.
b Introduced species.
e A, annual; P, perennial.
d Sp, spring; Su, summer; A, autumn; W, winter.
N.B. Animal matter found in the crops of Zebra Finches: termites (North 1909,
Immelmann 1962a); dipterans (Immelmann 1962a); heteropterans (Morton and
Davies 1983); aphids, ostracods, snails (Zann and Straw 1984a).
Appendix 3
Locomotory Ca lis
hopping Tet Call
walking Distance Call
pivoting Stack Call
flying Wsst Attack Call
taking-off Thuk can
landing Distress can
hovering Kackle can
Ark can
Recuperative Whine can
perching normal Begging can
perching fluffed Long Tonal can
squatting
sleeked-bill open Agonistic
carpel raising horizontal-sleeked
sleep-bill buried vertical-sleeked
sleep-head withdrawn jabbing
bill-fencing
Nutritive supplanting
pecking aerial combat
mandibulation chasing
Jumpmg
pulling Sexual
shovel-away Head-tail twist
foot -dam ping Upright fluffed
drinking allopreening
head-shake dumping
head-flick bill-wiping
defecating hop-pivot
hopping-to-and-fro
Exploratory Undirected Song
monocular gaze Directed Song
binocular gaze T ail-q ui vering
neck -stretch copulation
292 Appendix 3
Comfort Breeding
self-preening head-down tail-fan
bill-wiping nest ceremony
head-wiping pull-in building movement
yawnmg push-up building movement
wing -leg-stretch push-out-shake building movernent
wing-arch incubation
head-scratch begging head-up
head--up body-shake begging head-down
head-down body-shake (bathing) food regurgitation
tail-flick food transfer
wing-flick nest-leading
sun-bathing sprawl
Handbook summary of major biological parameters. Numerical values are means or medians and sources may be
found in the texto W: free-living Australian Zebra Finches, WC: wild-caught Australian Zebra Finches; D: domesti-
cated Australian Zebra Finches; L: captive-bred Lesser Sundas Zebra Pinches from Timor.
Classification
Family Passeridae (Vigors, 1825)
Subfamily Estrildinae (Blyth, 1829)
Tribe Poephilini
Genus Taeniopygia (Reichenbach, 1862)
Species guttata (Vieillot, 1817)
Subspecies guttata (Vieillot, 1817), Lesser Sundas
castanotis (Gould, 1837), Australia
Captive breeding
Recommended density per pair 0.3-0.4 m 3 per pair
Minimum temperature >12°C
Preferred temperature range 20-30°C
Breeding diet dry seed, soaked seed, seeding grasses, greens,
cuttle-bone, grit (siliceous and calcareous),
egg food, vitamin supplements.
Seed preferences 1. panicum, 2. white millet, 3. Japanese
millet, 4. canary seed
N
\.D
..¡:;.
Captive breeding (continued) ~
30 (D) ~
Colour morph varieties ~
('\:)
Hybrids (Immelmann, 1962a) Double-barred Finch ~
~
Long-tailed Finch ~.
Black-throated Finch ..¡:;.
Masked Finch
Star Finch
Plum-headed Finch
Diamond Firetail
Chestnut-breasted Finch
Bengalese Finch
African Silverbill
Indian Sílverbill
Grey-headed Silverbill
Tri-coloured Munía
Handling
Wild capture mistnets around nests or
water, walk-in traps
Marking metal bands (size) 02-Lambournes, D-Hughes
colour bands (size) XCS-Hughes
colour attractive to females red (WC, D)
colour unattractive to females light green (WC, D)
colour attractive to males black (D)
colour unattractive to males light blue (D)
Anaesthesia (50:50 mixture injected L M.)
ketamine (mg/g) 0.03
xylazine (mg/g) 0.006
Morphology
Mass (g) male 11.9-12.4 (W), 10.4-11.8 (L), 12.7 (D)
female 12.2-12.7 (W), 10.0-10.3 (L), 13.0 (D)
Wing length (mm) male 55 (W), 52 (L), 56-57 (D)
female 55 (W), 52 (L), 56-57 (D)
Billlength (mm) male 10.0 (W), 9.0 (L), 10.4 (D)
female 10.0 (W), 9.0 (L), 10.6 (D)
Bill depth (mm) males 8-9 (W), 7 (L), 6.5 (D)
female 7-9 (W), 7 (L), 6.6 (D)
Head-billlength (mm) males 23.3 (W)
female 23.6(W)
Rectrices length (mm) male 35-38 (W), 34 (D)
female 35-38 (W), 34 (D)
Remiges length numbers 1-9 (mm) male 40-46 (W)
female 41-46 (W)
Bill colour (Munsell scores) male 8.75,4/12 (W, D), 10,4/12 (L)
female 10,5/12 (W, D), 10,5/12 (L)
Iris colour (Munsell scores) male 10,3/10 (W), 10,3/1, 12,3/6 (D)
female 10,3/10 (W), 10,3/1, 12,3/6 (D)
Foot colour (Munsell scores) male 10,5/8 (W), 10,5/8, 12.5,5/10 (D)
female 10,5/8 (W), 10,5/8, 12.5,5/10 (D)
Back colour (Munsell scores) male 20,4/3 (W), 20, 4/1 (D) ~
~
female 20, 4/3(W), 20, 4/1 (D) ~
~
Tarsus length (mm) male 14.3-15.0 (W), 14.8 (D) ;::¡
~
female 14.3-15. O (W), 14.8 (D) ~.
Primaries 9 ..¡::.
Secondaries 9 N
T estis length (mm) male 14-25 (W) \O
V,
N
\O
0\
~
Morphology (continued) '\::l-
'\::l-
Testis length (mm)-dehydrated male 2-3 (WC) (':l
~
Testis mass (mg) male 1-51 (W) k
~.
Testis mas s (mg)-dehydrated male 12-41 (WC) ..¡::.
Ovary mas s (mg) female 20-170 (WC)
Ovary mass (mg)-dehydrated female 2-37 (W)
Moult
Onset-sex plumage (age in days) male 35-40 (W, D)
Completion-sex plumage (age) male 55-70 (W, D)
Onset-primary moult (age) male 80 (W)
female 80 (W)
Primary mouIt cycle (days) male 230-287(W)
female 240-287 (W)
Primary moult type continuous step-wise/serially descendent
Nutrition
Seed (g) required per day at d5°C 5 (WC)
20-30°C 3-4 (WC, D)
>30°C 1-2 (D)
Seed (g) required per day to feed five fledglings 14 (D)
Energy intake (kJ per day) at 2rc 35.7 (D)
Standard metabolic rate (k] per hour) at 23°C 0.80-0.88 (D)
Digestive efficiency (%) 78-88 (D)
Maximum seed intake (g per hour) 1-3.5 (WC), 3-6 (D)
Ingestion to egestion (minutes) at 23°C 87-106 (D)
Crop capacity (g) 0.3-0.6 (D)
Water intake (mI per day) at 23°C 3.1-4.6 (WC, D)
> 40°C 6-12 (WC, D)
Reproduction
Spermatogenesis-onset (days post-hatch) 60 (WC), 70 (D)
Ovulation-onset (days post-hatch) 60 (WC), 90 (D)
Pair formation-onset (days post-hatch) 50-60 (W,D)
Pair formation time (days) 2-10 (W, D)
Nest construction time (days) 5-13 (W), 1-5 (D)
Fertile period (days) 14-15 (D)
Onset of fertility (days before first egg) 11 (D)
Termination of fertility day before last egg (D)
Fertility (% eggs hatched) 75-96 (W)
Extra-pair paternity (% young) 2.4 (W), 5.6 (D)
Extra-pair paternity (% broods) 8 (W), 11 (D)
Brood parasitism (% broods) 13-32 (W)
Laying time (diurnal) first two hours of daylight (W, D)
Laying interval (h) 24 (W,D)
Clutch size (mode) 5 (W), 4-6 (D),
Laying type indeterminate
Egg mass (g) 1.0 (W, D)
~
Egg volume (mm3 ) 927 (W), 948 (D) '1:j-
'1:j-
Egg composition ~
;;'!l
water content (% mass) 76.4 (W), 75.3 (D) ~
~.
shell (% wet mass) 9.1 (W), 8.4 (D) .¡:.,
Calcium (mg) 18 (D)
N
\O
""-1
N
\O
Reproduction (continued) 00
yolk content (% wet mass) 26.4 (W), 24.2 (D) ~
lipid content (% wet mass) 6.9 (W), 5.8 (D) '"1:j-
'"1:j-
Onset of incubation (laying day) 4 (W), 1-4 (D) (':>
;:¡
37-92 (W), 20 (D) ~
Incubation bout (minutes) ~.
Incubatíon period (days) 11-14 (W), 11-14 (D) ...¡,.
Incubation temperature (oC) 36 (W), 35-37 (D)
Hatching interval (days) 1-2 (W), 2-5 (D)
Hatching mass (g) 0.6-0.9 (D)
Eyes open (days post-hatch) 6-7 (W, D)
Feathering complete (days post-hatch) 14 (W,D)
Nestling mortality (%) 9 (W), 13 (D)
Fledging mas s (g) 10 (W), 10 (D)
Age at fledging (days post-hatch) 16-18 (W), 17-22 (D)
Inter-brood interval (days) 52 (W), 57 (D)
Nutritional independence (days post-hatch) 35 (W)
Roosting independence (days post-hatch) 50 (W)
Successive broods >3 (W), 19 (WC), 14 (D)
Population
Colony size (nesting pairs) 7-47 (W)
Colony density (nests per hectare) 0.7-76.0 (W)
Population size <200-18,000 (W)
Recapture rate (years to 10% threshold) 0.1-1.6
Mortality: annual adult (%) 72-96 (W)
nestling-independence (%) 67(W)
nestling-sexual maturity (%) 77 (W)
first 12 months of life (%) 96 (W)
Life expectancy at hatching (days) 51 (W)
Maximum lifespan (years) 3.0-5.6 (W), 5-7 (D)
Sex ratio-adults (% males) 51-53 (W), 47-61(D)
-young (% males at 40 days) 44-52 (W)
Breeding philopatry (% natal pairs) 22(W)
Nesting predation (% lost) 66(W)
Rearing success (% nestlings fledged) 87-100 (W)
Breeding success (> 1 fledgling)
(% breeding attempts) 35(W)
(% incubated eggs) 41 (W)
(% clutches) 44-74 (W)
Age classes (days post-hatch)
Nestling 1-17
Juvenile 18-35
y oung immature 36-50
Late immature 51-80
Young adult 81-100
Adult >100
Temperatures
Diurnal cloacal temperature (oC) 38--44 (WC, D)
Critical temperature (oC) 45--46 (WC, D)
Thermal neutral zone (oC) 30--40 (WC, D)
Thermal conductance (j-g-h-oC) 5.2 (D)
Parasites ectoparasites mites, mallophagan lice, hippoboscid flies
endoparasites nematodes :;t..
~
Behaviour ~
(1:>
~
Subsong (days post-hatch) 28-50 (W), 24-50(D) ~
~.
Plastic song (days post-hatch) 50-80 (W,D)
~
Full song (days post-hatch) 80 (W,D)
N
\O
\O
w
o
o
~
Behaviour (continued) "'(j-
"'(j-
Number of distinct calls 12 (W, D), 4+ (L) (1:)
~
Main frequency of vocalisations (kHz) 2-5 ~
~.
Frequency of maximum emphasis (kHz) 3.6 (W, D), 4.7 (L) ..¡:..
Hearing sensitivity (kHz) range 1-6 (D)
best frequency 4(D)
Temporal auditory acuity (ms) 2.5-7(D)
Sensitive phase (days post-hatch)
song stage I 18-35 (D)
stage II 36-65 (W, D)
imprinting stage I 15-40 (D)
stage II >90 (D)
References
Wallace, A. R. xiii
Sargent, T. D. 168 Walter, H. 21, 115
Schleucher, E. 62, 66, 70 WaIter, M. J. 250
SchIinger, B. A. 243 Walters, M. J. 223, 244
Schmidt-Nielsen, K. 53, 67-9, 278 Webb, D. R. 96
Schodde,R.2, 7-8,11,13,60,153 Weisman, R. 251-2, 258
Schoepfer, M. 40, 281 White, C. M. N. 10,15-17
Schopfer, M. 49, 135 White, S. R. 27
Schubert, C. A. 84,261 Whitlock, F. L. 75
Schwab, C. 231 Whittell, H. M. xiv
Sedgwick, E. H. 29 Wiens, J. A. 46, 47
328 Author index
Williams, H. 194,214,219,224,228,231, Yoneda, T. 206
233,239-42,246
Willson, M. F. 49 Zann, R. 24,36-9,44-6,48,53,55,73-7,
Winfield, C. xiv 82-6,90-2,95-100,105,108-9,117-18,
Winkworth, R. E. 115 120-1,125, 128, 130-7, 140-50, 164,
Woinarski, J. c. Z. 40,145,281 170,174-5,182,187,193,197,201,
Wolters, H. E. 2, 102 205-8,210,214--16,219-21,230-2,234,
Workman, L. 172 259,265-7,283,285
Wyndham, E. 41, 49,125-6 Ziswiller, V. 2,5,53,97, 103
Wynn, S. E. 261 Zweers, G. A. 60,61,62,279
Subject index
Undirected Song
contexts 166, 168, 225
young
hormonal control 223
courtship and copulation in 189
performance 222
posture 222-3 development of independence 190-2
flocks 59, 189
possible functions 225
seed-dehusking development 190
Uraeginthus granatina 279,281