Breeding biology and nest site characteristics of the Rosy-faced Lovebird Agapornis roseicollis in Namibia

Abstract

The breeding biology of the Rosy-faced Lovebird Agapornis roseicollis was investigated in its natural habitat at three Namibian localities: Claratal, Hohewarte and Haris. The lovebird nests in colonies often shared by Sociable Weavers Philetairus socius. Birds nested in trees at a mean height of 3.8m, on telephone poles at 6.6m, windmills at 1.2m, and artificial nest boxes at 3.3m. Acacia erioloba and A. karroo were most often used for nest location. Nest tree habitats had low density vegetation with short (4m) trees, mostly A.. erioloba, spaced at distances of about 10m. No specific nest entrance orientation was chosen. Birds obtained nest materials from the bark of small branches, branchlets from the tips of branches, twigs, sticks, leaves and thorns of trees, predominantly A. karroo, A. erioloba, Ziziphus mucronata and Boscia albitrunca. Nine colonies, comprising 20 nests, were monitored every four days to establish incubation and fledging periods. Rearing
and fledging of chicks was found to be successful in eight nests. Measurements of 18 young from four nests were used to monitor growth rate. Nesting success was calculated at 0.1, following the revised Mayfield method. Mean clutch size at laying, hatching, and fledging was 4.42 ± 1.7, 2.26 ± 2.1 and 1.65 ± 2.1 (n = 20), respectively. There was no significant difference in mean mass, nor bill and tarsus length of young that hatched first or subsequently, but a Duncan test revealed a significant difference in mean mass of young that hatched first or fifth.

Full text

Copyright © NISC Pty Ltd
OSTRICH
EISSN 1727–947X
doi: 10.2989/OSTRICH.2007.78.1.3.47
Ostrich 2007, 78(1): 13–20
Printed in South Africa — All rights reserved
Breeding biology and nest site characteristics of the Rosy-faced Lovebird
Agapornis roseicollis in Namibia
Henry Ndithia1, Michael R Perrin1* and Matthias Waltert2
1Research Centre for African Parrot Conservation, School of Biological and Conservation Sciences,
University of KwaZulu-Natal, Private Bag X01, Scottsville 3209, South Africa
2Centre for Nature Conservation (Department 1), Georg-August-Universität Göttingen, Von-Siebold-Straße 2, 37075 Göttingen, Germany
* Corresponding author, e-mail: perrin@ukzn.ac.za
Introduction
African lovebirds are the only parrots, other than Monk
Parakeets Myiopsitta monachus, that construct their own
nests. Four species (Agapornis personata,A. fisheri, A.
lilianae and A. nigrigenis) build domed nests, whereas A.
roseicollis builds a cup-shaped nest within cavities (Eber-
hard 1998). In Agapornis species, nest building is asso-
ciated with colonial breeding and nest building in parrots
(Psittaciformes) either evolved from the habit of lining the
cavity with nest material, or through nest site adoption.
The Rosy-faced Lovebird is a non-territorial, sociable and
gregarious species (Forshaw 1989) that is highly water-
dependent (Harrison et al. 1997). They roost and nest in
shrub grassland where they may feed, although most feed-
ing occurs in highly wooded riparian vegetation on dry
water courses. The Rosy-faced Lovebird is sedentary (Fry
et al. 1988), occurring largely at the Namibian escarpment
(Harrison et al. 1997) and is near-endemic to Namibia, with
a range exceeding 50 000km2(Simmons et al. 2001).
Rosy-faced lovebirds are colonial breeders with natural
breeding sites in the inaccessible and often vertical cracks
found in steep rock faces on exfoliating granite or sandstone
koppies (Harrison et al. 1997).
However, Agapornis roseicollis is highly adaptable and
nests in communal nests of the Sociable Weaver Philetairus
socius and White-browed Sparrow Weaver Plocepasser
mahali (Fry et al. 1988) and on telephone poles (Harrison et
al. 1997). Nest site selection is a key determinant of repro-
ductive success; sites offer protection from predators and
harsh environmental conditions. Peak egg-laying months
are February–March (Harrison et al. 1997).
The viability of an avian population depends on the deli-
cate balance between natality and mortality (Johnson
1979), and we infer demographic status by estimating rates
of births and deaths. This is not always easy, but the
proportion of eggs that hatch can be used to gauge natality
as a measure of reproductive performance.
Here we present original breeding data from a wild popu-
lation, since previous information on the species’ breeding
biology was based on captive studies. Since most nests
were found during the incubation and the nestling periods, it
was difficult to estimate when nest building had begun. The
objective of this study was to investigate the nesting char-
acteristics and breeding biology of the Rosy-faced Lovebird
in its natural habitat and to identify plant species critical for
its reproductive success.
Study area and methods
Fieldwork was carried out from January to July 2004 as part
of a wider study on the ecology and status of the Rosy-
faced Lovebird. The study was based at three Namibian
localities: Claratal (22°44’S, 16°45’E), Haris (22°43’S,
16°51’E) and Hohewarte (22°37’S, 17°20’E), in the Nami-
bian escarpment at altitudes between 1 778m and 2 114m
above sea level. The localities contained dense shrub
savanna and scattered trees; however, grassland with
The breeding biology of the Rosy-faced Lovebird Agapornis roseicollis was investigated in its natural habitat at three
Namibian localities: Claratal, Hohewarte and Haris. The lovebird nests in colonies often shared by Sociable Weavers
Philetairus socius. Birds nested in trees at a mean height of 3.8m, on telephone poles at 6.6m, windmills at 11.2m, and artifi-
cial nest boxes at 3.3m. Acacia erioloba and A. karroo were most often used for nest location. Nest tree habitats had low
density vegetation with short (4m) trees, mostly A. erioloba, spaced at distances of about 10m. No specific nest entrance
orientation was chosen. Birds obtained nest materials from the bark of small branches, branchlets from the tips of branches,
twigs, sticks, leaves and thorns of trees, predominantly A. karroo, A. erioloba, Ziziphus mucronata and Boscia albitrunca.
Nine colonies, comprising 20 nests, were monitored every four days to establish incubation and fledging periods. Rearing
and fledging of chicks was found to be successful in eight nests. Measurements of 18 young from four nests were used to
monitor growth rate. Nesting success was calculated at 0.1, following the revised Mayfield method. Mean clutch size at
laying, hatching, and fledging was 4.42 ± 1.7, 2.26 ± 2.1 and 1.65 ± 2.1 (n = 20), respectively. There was no significant differ-
ence in mean mass, nor bill and tarsus length of young that hatched first or subsequently, but a Duncan test revealed a
significant difference in mean mass of young that hatched first or fifth.

Ndithia, Perrin and Waltert14
26–50% scattered shrub cover was the dominant vegeta-
tion. The study sites were located on commercial farms
managing domestic animals (mainly cattle) and wildlife.
Water was provided in many localities on the farms and
was readily available to the lovebirds. Nest searching
involved walking the plains and driving along gravel roads
next to farms. Once discovered, the positions (location,
elevation and distances between sites) of the nest colonies
were recorded using a Garmin eTrex Venture GPS. Nest
contents were monitored.
Heights of nests from ground level and distances
between nest trees were measured. Ladders were used to
access nests in trees, or an elliptical mirror — attached to
an easily foldable aluminum pole — was used to examine
the nest contents and the internal shapes of the nests.
Information on nesting site characteristics was collected,
including the plant species on which the nest was located,
height of the tree, height of the nest above ground, eleva-
tion (height above sea level), co-ordinates at nest locations,
nest orientation, and habitat type.
Observations on breeding biology were made simultane-
ously with observations on feeding behaviour. Nest monitor-
ing was done early in the morning (06:00–07:00 in summer
and 06:45–07:00 in winter). Most of the nests in Hohewarte
were in nest boxes with small nest entrances. Hence, nests
were monitored with the aid of a mirror and a torch. At
hatching, and at four-day intervals, the following measure-
ments were taken: date, nest number, colony, age of chicks
(in days), body mass and wing, bill and tarsus lengths.
Nesting success was estimated using the method of
Mayfield (1975), namely: success = (1 – [losses/exposure])
^np where exposure is the total number of active nest days,
and np is the period in the nesting cycle being considered.
Overall, nesting success was calculated as the product of
success during the laying and incubation periods combined,
and the nestling period. The variance of Mayfield’s estima-
tor was derived from the following expression, developed
by Johnson (1979): ([exposure – losses] Xlosses)/([expo-
sure] ^3).
Breeding behaviours were observed using Bushnell 8x42
binoculars and a Kowa 40Xtelescope. Activity and behav-
iour of the flock, time, date and flock size were recorded.
Rosy-faced Lovebirds are sexually monomorphic (Forshaw
1989) and pair bonds were only assumed when individuals
performed certain reproduction-related behaviours.
Eighteen young were marked with either plastic colour
bands, permanent ink or nail varnish, to distinguish them
individually.
Results
Breeding behaviour
Nest locations
In the study locations, lovebirds nested in communal nests
with Sociable Weavers. Nine colonies, comprising 20 nests,
were active. Nests were observed in Acacia spp. trees, on
telephone poles, on windmill structures, and in artificial nest
boxes. Nests were located in trees at a mean height of 3.8m
above ground, at 6.6m on telephone poles, at 11.2m on the
windmill and at 3.3m on nest boxes above ground (Table 1).
At Claratal, birds preferred to nest in dead and/or partially
dead trees, but five nests located in trees failed (Figure 1).
Two of six nests located on telephone poles, and five of
eight nests in nest boxes, were successful. Most of the nests
were found on structures other than trees. Birds added nest
materials to the nests during and after incubation.
Choice of nest trees
The lovebirds nested exclusively in Camel Thorn Acacia
erioloba or Sweet Thorn Acacia karroo. A. erioloba was the
most preferred tree species for nest location (71%), perhaps
because these trees also served as a source of food and
nest-building material. A. erioloba occurred mainly in the
drier parts of the lovebird’s habitat, while A. karroo occurred
in the open grassland and along dry river courses.
Nearest neighbour for nest site trees
Distances from the nest tree to the nearest adjacent tree
were small, ranging from 1.6–20m (Table 2). This, however,
was a poor indicator of the density of the vegetation around
nest trees because most plants found near nest trees were
n Mean ± SD
Trees
Tree height 5 5.9 ± 1.0
Nest height 5 3.8 ± 0.8
Elevation 5 1 929.2 ± 6.2
Telephone poles
Telephone pole height 6 7.8 ± 1.8
Nest height 6 6.6 ± 0.6
Elevation 6 1 962.3 ± 26.4
Nest boxes
Structure height 8 3.5 ± 0
Nest height 8 3.3 ± 0
Elevation 8 1 793.0 ± 0
Windmill structures
Windmill height 1 14.0
Nest height 1 11.2
Elevation 1 1 928.0
Table 1: Mean height above ground (m) of nests and nest struc-
tures and mean elevation (m above sea level)
HOHEWARTE HARIS
NUMBER OF NESTS
Tree spp. Windmill Telephone
pole
Human
structure
Telephone
pole
1
2
3
4
5
6
7
8
CLARATAL
Figure 1: Nest success of Rosyfaced Lovebirds according to nest
site and study location; black = total number of nests, white =
success, grey = failure

Ostrich 2007, 78(1): 13–20 15
very short, ranging from 1.5–7m in height, and did not offer
effective cover to the nest trees. Nest trees were conspicu-
ous among smaller trees and shrubs.
Mean tree height for the three locations was not a good
indicator of the environment around the nest trees because
eight nests, mostly in nest boxes, in Hohewarte, formed
one colony, with its neighbouring tree 15m high. This was
not characteristic of the other two study sites where mean
height of the 12 nests there was much lower (4m).
A. erioloba was the commonest species adjacent to the
nest trees and structures, with a higher frequency than A.
karroo and A. mellifera (Figure 2). These trees provided
food and cover for the lovebirds.
Nest entrance orientation
Nest entrance orientation was predominantly westerly
(Figure 3). Eleven nests (55%) had their entrances facing
north-west, four to the south-west and two to the west. The
predominant wind direction (north and north-west) probably
affected the direction of nest entrances. Most nest
entrances faced into the wind.
Nest building and building materials
Nest building materials were obtained from the bark of
branches, branchlets obtained from the tips of branches,
twigs, sticks, leaves, and thorns. The branchlets, leaves
and thorns were mainly accessed from A. karroo and A.
erioloba. Buffalo-thorn Ziziphus mucronata, Shepherd’s
Tree Boscia albitrunca,Acacia erioloba, and A. karroo
were the sources of strips of bark, which were the most
commonly-used nest-building material. These materials
were used to line the inside of the nests, especially where
the eggs were laid and where nestlings rested. Nest mate-
rials were used to block nest entrances so that eggs and
young would not fall from the nest.
Nest materials were cut using the strong, sharp bill and
transported under the feathers of the rump and the flanks.
Grasses were not collected, nor used to line the nest.
Birds made two cuts on the bark of a small branch, one
above and one below the patch to be stripped. They then
used the bill to peel off the bark in narrow strips, tucking
them under their feathers. Birds flew to the nests and
inserted the materials at the sides of the nest, leaving a
medium-sized, elliptical cup-shaped nest at the centre.
The Sociable Weaver nest colony was made of grass
woven together to form a huge neatly-woven structure.
The communal nests of Rosy-faced Lovebirds and
Sociable Weavers were located within this structure.
Each communal nest colony contained many breeding
chambers, sometimes connected to one another, espe-
cially those occupied by lovebirds. There were two open-
ings to each nest, one for entry and the other for escape. It
is not known whether previously-used nests were reused,
but it is likely that they were.
Breeding biology
Seasonality
Eighteen colonies, comprising 32 nests in three locations,
were identified between January and May 2004. Of these,
nine colonies with 20 nests were active. Nest colonies of A.
roseicollis occurred communally with those of Sociable
Weavers Philetairus socius. They were permanent struc-
tures, and were likely to be used for several breeding
seasons.
The timing of laying and incubation by different pairs
was closely synchronised with each other and with rain-
fall (Table 3), photoperiod and food availability. Laying
and incubation at the three study locations ran from
February through to the end of March, during the peak of
the rainy season, when grass seeds were abundant.
Description n Mean ± SD Min. Max.
Distance 20 10.2 ± 8.4 1.6 20
Height of nest tree/structure (at three sites) 20 8.4 ± 5.7 1.5 15
Height of nest tree/structure (at two sites) 12 4.0 ± 2.0 1.5 7
Table 2: Distances (m) from nest trees or nest structures to nearest neighbouring trees or shrubs, and heights (m) of nest trees/structures
SPECIES
FREQUENCY
Acka Acme Acer Lyvi
2
4
6
10
8
Figure 2: Frequency with which species were nearest neighbour to
nest trees or structures. Acka = Acacia karoo, Acme = A. mellifera,
Acer = A. erioloba, Lyvi = Lycium villosum
FREQUENCY
2
4
6
10
8
NW SW W E SE NE
DIRECTION
Figure 3: Frequency of nest entrance orientations

Ndithia, Perrin and Waltert16
Rains gradually declined during April and ceased in May.
When hatching began in June, nestlings appeared and fledg-
ing occurred. Food, however, was still available. Anthephora
schinzii, a preferred grass species, had by then ‘dropped’ all
seed from its inflorescences and birds shifted attention to
other food items, but continued to feed on seeds of this
grass species from the ground (Ndithia and Perrin 2006a).
Clutch size and breeding success
Clutch size varied from 3–8, with a mean of 4.4 ± 1.7 (n =
19) (Table 4). There was no significant difference in clutch
size between the different locations (one way ANOVA, p =
0.19 > 0.05), but there was a marked difference in clutch size
between laying, hatching and fledging (Tukey HSD for
unequal N, p = 0.00055 < 0.05; Table 5).
Eggs were laid at intervals of one day; that is, one egg
every other day. Incubation began with the first egg laid, so
the young hatched asynchronously, causing a considerable
difference in size of individuals of the same brood. Incubation
and fledging lasted 23 and 43 days, respectively (Table 6).
A total of 711 nest days of observation were used to
determine breeding success: 337 during the incubation
period and 374 during the nestling period. Seven nests
were lost during incubation and 13 during the nestling
period. Extrapolation determined the onset of incubation for
nests that were well advanced.
Breeding success was computed using the Mayfield
method (Table 7). The daily survival rate of chicks during
incubation and the nestling period were 0.97 and 0.97,
respectively, predicting survival rates of 0.62 and 0.23 for
these developmental periods. This generates an overall
survival rate of 0.14 from egg laying to fledging, and 1.65
successful fledgings per successful nest.
Growth rate and development
Four nests, comprising 18 young, were monitored. The fifth-
hatched chick was smaller than the previously-hatched
chicks (Table 8). Mean bill and tarsus length of chicks
decreased slightly from early to late chicks (Table 9).
One-way ANOVA did not reveal any significant differ-
ences in mean body mass, or bill and tarsus length of
successively-raised young. However, a post hoc ANOVA
with a Duncan test revealed a significant difference in
mean body mass of the young hatched first and hatched
fifth (p = 0.0370 < 0.05; Table 10). A Duncan test did not
reveal any significant difference between the mean bill or
tarsus length of young that hatched first or fifth. Correlation
analysis showed that the three variables body mass, bill
length and tarsus length were significantly correlated
(Table 11). Table 12 summarises the relationship between
the parameters (y = –43.5 + 6.3 b + 0.95 t, where y =
mass, b = bill and t = tarsus).
All pairs raised a single brood and the pattern of feather
development of chicks was uniform. Upon hatching, young
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
71.1 73.7 76.2 50.8 12.7 0 0 0 6.4 12.7 38.1 50.8
Table 3: Mean monthly rainfall (mm) for 2003 for Windhoek, Namibia, near the study site
Parameter Location
Claratal Haris Hohewarte Total
Laying 4.13 ± 1.7 (8) 4.33 ± 1.1 (3) 4.75 ± 2.0 (8) 4.42 ± 1.7 (19)
Hatching 1.63 ± 2.3 (8) 3.33 ± 0.6 (3) 2.05 ± 2.1 (8) 2.26 ± 2.1 (19)
Fledging 1.11 ± 2.2 (9) 1.00 ± 1.7 (3) 2.50 ± 2.1 (8) 1.65 ± 2.1 (20)
Table 4: Mean clutch size (± SD (n)) at laying, hatching and fledging at three different locations
Laying (1) Hatching (2)
M = 4.2000 M = 4.2000
Laying (1)
Hatching (2) 0.007454*
Fledging (3) 0.000819* 0.724760
Table 5: Tukey HSD for unequal n test showing marked differ-
ences in clutch sizes at laying, hatching and fledging; *indicates
significant difference
Days n
Incubation
Claratal 23.3 3
Hohewarte 22.2 5
Haris 23.3 3
Mean = 22.9 Total = 11
Fledging
Claratal 42 2
Hohewarte 43.5 4
Haris 44 1
Mean = 43.2 Total = 7
Table 6: Mean incubation period and fledging time (days) for
Agapornis roseicollis at three study locations

Ostrich 2007, 78(1): 13–20 17
were covered with a rather dense reddish down. They grew
rapidly and after nine days pin feathers began to replace
the natal down. The eyes opened about 11 days after
hatching, and the parents fed the young by regurgitation.
After 13 days, the bodies of young were entirely covered
with down and pin feathers. After 18 days, green juvenile
plumage started to appear on the wings and body, and the
bill was well developed and horny-black. At 22 days, the
young were fully feathered (wings, head and tail) with dull
green plumage that covered the grey down beneath.
At Day 27, pale orange feathers had developed on the neck
and breast. At Day 34, more pale orange feathers had devel-
oped; the rump was blue and the tail had red tips on the outer
tail feathers. Adult plumage was almost fully developed, but
the neck and breast were still a pale orange colour, in contrast
to the adult plumage of bright red to rose pink.. The upper half
of the bill was black, the lower half dark yellow, in contrast to
the pink of the adult bill. Young attempted to fly at Day 39, but
had poor flight capabilities. Young fledged at 42–44 days and
flew well. They made short flights from the nests, and flew
back to the nest at any slight disturbances, and did not fly
again for several hours.
Discussion
Nest site selection
There was a relatively high success rate of nests located on
telephone poles, although two had dead chicks, not caused
by predation but perhaps attributable to disease or other
factors. Nests in nest boxes had a high success rate, prob-
ably because of their location adjacent to human habitation
where predators were absent. Lovebirds preferred struc-
tures to trees as nest site locations because they were
higher above ground and, most probably, in more secure
situations. Snakes easily climbed trees to reach nests but
were less likely to climb a structure like a wall or a pole or
iron structures. Birds avoided bushes, and nest colonies on
trees were usually on, or very near, the main trunk, high in
the canopy. Alonso and Muñoz-pulido (1991), in a study of
nest site selection and nesting success of the Azure-winged
Magpie Cyanopica cyana, showed that nesting success
was not correlated with nest height above the ground.
Santisteban et al. (2002) showed that elevated nest height
increased nest vulnerability to visually-oriented nest preda-
tors in the Fish Crow Corvus assifragus.
Nest re-use
Whether a pair uses the same nest and/or colonies in the
subsequent breeding seasons remains unknown. The same
nest colonies were apparently re-used in subsequent
breeding seasons, as lovebirds were observed entering old
nests, but some Sociable Weaver colonies had no active
(1) (2) (3) (4) {5}
Young M = 48.4 M = 46.3 M = 45.6 M = 40.3 M =34.1
1 (1) 0.728 0.668 0.232 0.037
2 (2) 0.912 0.360 0.070
3 (3) 0.387 0.077
4 (4) 0.312
5 (5)
Table 10: Duncan test showing significant difference at p < 0.05 of the mean of body mass of young that hatched first and fifth. Note that p =
0.037 for means of Young 1 and 5
Variable Body mass Bill Tarsus
Mass 1.00
Bill 0.94 1.00
Tarsus 0.86 0.89 1.00
Table 11: Correlations of body mass, and bill and tarsus length. All
correlations are significant at p < 0.05 (n = 83)
Beta se Beta se t(80) p-level
Intercept -43.46 4.08 -10.64 0.00
Bill 0.79 0.085 6.31 0.67 9.37 0.00
Tarsus 0.16 0.085 0.95 0.51 1.85 0.07
Table 12: Summary of the regressions for body mass, bill and
tarsus length. Adjusted r² = 0.8765
Incubation Nestling Hatching Overall
0.62 0.23 0.51 0.14
Table 7: Nest success probabilities averaged for all localities
Hatched Mean ± SD n
First 48.4 ± 15.1 19
Second 46.3 ± 14.9 20
Third 45.6 ± 17.2 20
Fourth 40.3 ± 21.4 17
Fifth 34.1 ± 23.0 16
Table 8: Mean body mass of first-, second-, third-, fourth- and fifth-
hatched young across all clutches
Bill Tarsus
First 12.1 ± 1.7 (18) 17.6 ± 1.7 (18)
Second 12.0 ± 1.7 (19) 17.8 ± 2.2 (19)
Third 12.0 ± 2.0 (19) 17.6 ± 2.7 (19)
Fourth 11.0 ± 2.4 (16) 16.3 ± 3.6 (15)
Fifth 10.9 ± 2.3 (13) 16.3 ± 2.6 (12)
Table 9: Mean bill and tarsus length (± SD (n)) of the first-,
second-, third-, fourth- and fifth-hatched young across all clutches

Ndithia, Perrin and Waltert18
lovebird nests. Some lovebird nests in the colonies were
found to be inactive throughout the breeding season. They
had been previously used, but were not re-used, and the
entrances were found to be blocked. During the nestling
period, nests were regularly cleaned of the wastes of the
young but after the chicks fledged, nests contained heaps
of dry hardened excreta, which may prevent the re-use of
the nests.
Parrots are known to have a strong fidelity for nesting sites
and use a small fraction of potential habitat for breeding (Synder
et al. 1987, Robinet and Salas 1999). At Lake Naivasha, Kenya,
hybrid lovebirds (A. personatus XA. fisheri) permanently occupy
the same tree cavities throughout the year (Thompson
1987). The re-use of nests by Rosy-faced Lovebirds (and
Black-cheeked Lovebirds (Warburton 2003)) remains an
open question. Nest boxes were re-used at Hohewarte and
inactive nests were used for roosting during the day.
Sociable Weavers constructed the huge grass-thatched
nest structures that formed the nest colonies. They were
often seen transporting grass materials to nest colonies, but
the lovebirds only lined the inside of the nests to make the
internal environment suitable for eggs and chicks. The huge
nests were large colonies established over many years, and
had taken much effort, time and energy to construct.
Threat to nesting locations
Nesting sites for breeding lovebirds are not limiting in the
natural habitat, but may be limiting with respect to tele-
phone poles and other man-made structures. A. karroo and
A. erioloba are common as nest trees and are important for
the continued survival and viability of the lovebird popula-
tion in Namibia, particularly if the lovebirds are unable to
access nesting sites associated with human habitation,
where breeding success is greater.
Nest orientation
Lovebirds placed their nests on structures with regard to
wind direction, i.e. westerly; winds are strong in some
months of the year in Namibia. This orientation may be
influenced by protection from rain and direct temperature
effects, which can be scorching. Only one nest faced east.
Nesting success
Few published data are available for the nesting success of
wild African parrot populations (Wirminghaus et al. 2001,
Symes 2004, Warburton and Perrin 2005, Taylor and Perrin
2006). Nestling survival data are, however, available for the
Australian Galah Eolophus roseicollis, but vary between
seasons and study sites (Rowley 1990). The percentage
fledged range from 58–65% for broods of three to five.
Smallest and largest broods raised a lower percentage of
fledglings, which probably reflects chilling of isolated nest-
lings and high mortality of the last hatched chick, respec-
tively. Reproductive success also varied with rainfall and
food availability.
Breeding stimulation and number of broods raised
Lovebirds raised one brood per pair during the fieldwork
period. It is not known if they raised a second or more
broods after July. Dilger (1960) reported that young A. rose-
icollis become sexually mature at 37 days after fledging but
only behaved sexually after the post-juvenile moult was
complete, at about four months. His data were, however,
based on captive studies, the relevance of which in the wild
is questionable. Simmons (1997) indicated that breeding
season extension of Namibian birds may occur opportunis-
tically when conditions are favourable, after the summer
rains, and that breeding coincides with the timing of maxi-
mum food availability immediately following heavy rain. This
appeared to be so for the lovebirds in this study, as breed-
ing followed the heavy rains that occurred in January and
February, which caused an abundance of grass seeds, the
preferred diet of the lovebirds.
In the arid and semi-arid subtropics, where rainfall is
erratic and unpredictable, it is rainfall, through its influence
on food availability, that is the key proximate determinant of
both clutch size and the timing of breeding in many bird
species (Marchant 1960, Immelmann 1973, Boag and
Grant 1984, Lloyd 1999). In southern Africa, the importance
of rainfall in the timing of breeding seasons is well estab-
lished (Maclean 1970, Immelmann 1973, Lloyd 1999). In
this study, the study sites were commercial wildlife and live-
stock farms supplied with artificial watering points. Water
restriction therefore did not limit the lovebirds from drinking
and did not prevent breeding or raising a brood (or subse-
quent broods). This was because the rains, not drinking
water, determined the advent of good seed and insect
crops, which in turn determined breeding success. A.
fischeri X A. personatus hybrids at Lake Naivasha, Kenya,
breed at any time in the year, mainly because of the bird’s
ability to feed on maize, which is widely grown (under irriga-
tion) in the area throughout the year (Thompson 1987).
Where rains occur twice a year in Kenya and when farmers
grow the seed crop fed upon by this lovebird, they breed
year round (Dodman pers. comm.).
Rosy-faced Lovebirds are entirely vegetarian (HN and
MRP unpublished data) and whether a second brood is
attempted may depend on rainfall and other climatic
factors, through their ongoing influence on food availability.
During the present study no rains came at the end of the
first brood in May, and most plants that formed part of the
lovebird’s diet had lost their seeds. Seed pods had hard-
ened after drying and became difficult to crush. Plant leaves
remained available but were less preferred. If these factors
are significant, it is unlikely that a second brood was raised.
The lovebird’s dependency on water, with respect to habi-
tat choice, however, was very pronounced, and their distri-
bution was localised, apparently because of water availabil-
ity. Lovebird nesting activity was conspicuously absent from
locations where water was absent. Telemetric, feeding and
drinking studies done alongside the breeding project
revealed long flight patterns to water points and feeding
grounds (Ndithia and Perrin 2006b).
Developments of young
The last-hatched chick realised a lower final asymptotic
body mass, probably because sibling rivalry caused less
feeding by the parent or because of contest, not scramble,
competition for food. Hatching asynchrony is adaptive: large
chicks survive while small ones have greater mortality rates.

Ostrich 2007, 78(1): 13–20 19
Likely causes of nest failure
In addition to water and food availability, nest site location
influenced the choice of breeding sites. All nests located in
trees failed and predation was observed in at least two of
the nests. Protection against predators was an important
consideration and nesting success was higher for nests
located on telephone poles and in nest boxes than in
acacia woodland. One cause of nest failure was predation
by a Boomslang Dispholidus typus found inside a lovebird
nest 4.2m above ground. Boomslangs are good tree
climbers and after the snake was removed from the nest,
no eggs or young remained. After a few days, a neighbour-
ing Rosy-faced Lovebird nest with seven eggs was lost,
which was attributed to snake predation. Boomslangs have
been recorded as a nest predator of Red Bishop Euplectes
orix (Friedl and Klump 2000), Green Wood-Hoopoe Phoe-
niculus purpureus (du Plessus 1989), Water Thick-knee
(Water Dikkop) Burhinus vermiculatus (Hockey et al. in
press), and Cape Rockjumper Chaetops frenatus (Holmes
et al. 2002). If all nests in which eggs and nestlings were
missing were attributed to predation, then predation
reduced the reproductive success by 73%.
Implications for conservation
Rosy-faced Lovebirds are only abundant locally in
Namibia, in pockets of habitats forming a metapopulation.
There is no apparent immediate threat to the populations
nesting in rock faces (Harrison et al. 1997). However, with
only eight of the 20 nests investigated here succeeding,
and only 1.65 chicks being reared per nest, there is
concern as to whether this rate of reproduction can
support the metapopulation. Can a nest and breeding
success rate offset lifetime mortality? With the de-listing of
the species from Appendix II of CITES, one wonders what
would happen to the population if commercial trade was
reintroduced.
The current numerical and demographic status of the
Rosy-faced Lovebird is unknown, so any policy suggest-
ing trade should be resisted. Therefore, one questions
whether the revision of the CITES status should be imple-
mented. There is certainly a need to maintain population
demographic and capture trend data through long-term
field studies, to investigate at what age young in the wild
start to breed, and to determine what impact this has on
the viability of the species. Simmons (pers. comm.), as
Namibia’s ornithologist, was consulted by CITES; he
recommended the down-listing because of the lovebird’s
widespread distribution (probably larger than its historical
range because of all the new water points), and the ease
with which it is bred in captivity. Although the lovebirds in
this study population had low breeding success, this
effect may be offset by the fact that most of the popula-
tion breed in inaccessible cliffs that neither snakes nor
even baboons are likely to reach. Thus, breeding success
is probably much higher than indicated by the present
study. In addition, the birds are capable of multi-broods
while conditions are good. Last, there is no indication
from parrot breeders that populations have declined
(Simmons pers. comm.).
Recommendation
Since little is known of the biology of this species in the
wild, further studies are necessary to contribute to knowl-
edge that will act as a basis for decision-making. It is espe-
cially important to investigate whether pests and disease
contribute to the nest failure and death of young reported in
this study. Further research efforts should investigate
avenues for increasing breeding success in wild popula-
tions, exploring reasons for breeding failure in trees and
what these mean for the viability of the species.
Acknowledgements — We thank the South African National
Research Foundation and the University of KwaZulu-Natal for
research funding, DAAD (Deutscher Akademischer Austauch
Dierst, the German Academic Exchange Programme) for a student
bursary (to HN), and Prof Michael Meuhlenberg for academic and
research planning. Drs John Mendelsohn and Chris Brown are
thanked for acting as supervisors during the field study, and provid-
ing guidance and assistance.
References
Alonso JA and Muñoz-pulido R 1991. Nest-site selection and
nesting success in the Azure-winged Magpie, Cyanopica cyana
in Central Spain. Bird Study 38: 45–51
Boag PT and Grant PR 1984. Darwin’s finches (Geospiza) on Isla
Daphne Major Galapagos: breeding and feeding ecology in a
climatically variable environment. Ecological Monograph 54:
463–489
Dilger WC 1960. The comparative ethology of the African parrot
genus Agapornis. Zietschrift für Tierpsychologie 17: 649–685
du Plessus MA 1989. Behavioural Ecology of the Red-billed Wood-
hoepoe Phoeniculus purpureus in South Africa. PhD thesis,
University of Cape Town, South Africa
Eberhard JR 1998. Evolution of nest building behaviour in Agapor-
nis Parrots. The Auk 115: 455–464
Forshaw JM 1989. Parrots of the World (3rd revised edn). David
and Charles, Newton Abbot, London
Friedl TWP and Klump GM 2000. Nest and mate choice in the
Red Bishop Euplectes orix: female settlement rule. Behavioural
Ecology 11(4): 378–386
Fry CH, Stuart K and Urban E (eds) 1988. The Birds of Africa,
Vol. 3. Academic Press, London
Hockey PAR, Dean WRJ, Ryan PG and Maree S (eds) in press.
Roberts’ Birds of Southern Africa (7th edn). Black Eagle Publi-
shing, Cape Town, South Africa
Holmes RT, Frauenknecht BD and du Plessis MA 2002.
Breeding system of the Cape Rockjumper, a South African Fyn-
bos endemic. Condor 104(1): 188–192
Immelmann K 1973. Role of the environment in reproduction as a
source of predictive information. In: Farmer DS (ed) Breeding
Biology of Birds. pp 121–147. National Academy of Sciences,
Washington, DC
Johnson HD 1979. Estimating nest success: the Mayfield method
and an alternative. Auk 96(4): 651–661
Lloyd P 1999. Rainfall as a breeding stimulus and clutch size
determinant in South African arid zone birds. Ibis 141: 637–643
Low R 1994. Africa: still the vast unknown? In: Endangered
Parrots. Blandford, England
Maclean GL 1970. The breeding seasons of the birds of the south-
western Kalahari. Ostrich Supplement 8: 294–301
Marchant S 1960. The breeding of some SW Ecuadorian birds.
Ibis 102: 349–381

Mayfield FH 1975. Suggestions for calculating nest success.
Wilson Bulletin 87: 456–466
Ndithia H, Perrin MR. 2006a. Diet and foraging behaviour of the
Rosy-faced Lovebird Agapornis rosiecollis in Namibia. Ostrich 77
(1&2): 45–51.
Ndithia H, Perrin MR. 2006b. The spatial ecology of the Rosy-
faced Lovebird Agapornis rosiecollis in Namibia. Ostrich 77
(1&2): 52–57.
Robertson A, Simmons RE, Jarvis AM and Brown CJ 1995.
Can bird atlas data be used to estimate population size? A case
study using Namibian endemics. Biological Conservation 71:
87–95
Robinet O and Silas M 1999. Reproductive biology of the endan-
gered Ouvea Parakeet Eunymphicus cornutus uvaeensis. Ibis
141: 600–669
Rowley I 1990. Behavioural Ecology of the Galah Eolophus rose-
icollis in the Wheatbelt of Western Australia. Surrey Beatty &
Sons Pty Limited, Chipping Norton, NSW, Australia
Santisteban L, Sieving KE and Avery ML 2002. Use of sensory
cues by Fish Crows Corvus assifragus, preying on artificial birds’
nests. Journal of Avian Biology 33: 245–252
Simmons RE 1997. Rosy-faced Lovebird, Agapornis roseicollis. In:
Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ,
Parker V and Brown CJ (eds) The Atlas of Southern African Birds.
Vol. 1: Non-passerines. BirdLife South Africa, Johannesburg
Simmons RE, Boix-Hinzen C, Barnes K, Jarvis AM and Robertson
A2001. Namibia. In: Fishpool LDC and Evans MI (eds) Impor-
tant Bird Areas in Africa and Associated Islands: Priority Sites for
Conservation. pp 639–660. Pisces Publications and BirdLife
International (BirdLife Conservation Series No. 11), Newbury and
Cambridge, UK
Snyder PP, Wiley JW and Kepler CB 1987. The Parrots of
Luquillo: Natural History and Conservation of the Puerto Rican
Parrots. Western Foundation of Vertebrate Zoology, Los
Angeles
Symes CT and Perrin MR 2004. Breeding biology of the Grey-
headed Parrot Poicephalus fuscicollis sauhelicus in the wild.
Emu 104(1): 45–57
Taylor S and Perrin MR 2006. Aspects of breeding biology of the
Brown-headed Parrot Poicephalus cryptoxanthus in South Africa.
Ostrich 77(3&4): 225–228
Thompson JJ 1987. Lovebirds at Lake Naivasha. Swara 10(5):
11–12
Warburton LS 2003. The Ecology and Conservation Biology of the
Black-cheeked Lovebird Agapornis nigrigenis in Zambia. PhD
thesis, University of Natal, Pietermaritzburg, South Africa
Warburton LS and Perrin MR 2005. Nest-site characteristics and
breeding biology of the Black-cheeked Lovebird Agapornis nigri-
genis in Zambia. Ostrich 76(3&4): 162–174
Wirminghaus JO, Downs CT, Symes CT and Perrin MR 2001.
Breeding biology of the Cape Parrot Poicephalus robustus.
Ostrich 72(3&4): 159–164
Received April 2005, accepted September 2006
Editor: RE Simmons
20 Ndithia, Perrin and Waltert