ArticlePDF Available

Nesting biology and food habits of the endangered Sakalava Rail Amaurornis olivieri in the Mandrozo Protected Area, western Madagascar

  • The Peregrine Fund Madagascar
  • The Peregrine Fund, Madagascar Project

Abstract and Figures

We studied the nesting biology and food habits of the endangered and endemic Sakalava Rail Amaurornis olivieri from July to November 2015 in the Mandrozo Protected Area, western Madagascar. Three nesting pairs were observed and their nests were constructed in a dense mat of reeds Phragmites mauritianus and averaged 56.7 ± 15.3 cm above the water (n = 3 nests). Nests were built by both adults and it took 3 d on average to complete a nest (n = 2 nests). Thirteen matings were observed and lasted 4.1 s on average (n = 2 pairs). Average clutch size was three eggs (n = 2 nests). Both sexes incubated; the incubation period was 15–17 d (n = 2 nests). Both male and female participated in brooding and feeding the young, which remained for 3 d in the nest and became independent of their parents at 45 d of age. Based on 194 identified food items, the Sakalava Rail’s diet was composed of invertebrates: spiders (53.1%), insects (32%), crustaceans (10.8%) and molluscs (4.1%). The home ranges of two radio-tagged individuals were 0.95 and 1.98 ha.
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
Journal of African Ornithology
ISSN: 0030-6525 (Print) 1727-947X (Online) Journal homepage:
Nesting biology and food habits of the endangered
Sakalava Rail Amaurornis olivieri in the Mandrozo
Protected Area, western Madagascar
Yverlin Z Pruvot, Lily-Arison René de Roland, Gilbert Razafimanjato, Marius
PH Rakotondratsima, Aristide Andrianarimisa & Russell Thorstrom
To cite this article: Yverlin Z Pruvot, Lily-Arison René de Roland, Gilbert Razafimanjato, Marius
PH Rakotondratsima, Aristide Andrianarimisa & Russell Thorstrom (2018) Nesting biology and
food habits of the endangered Sakalava Rail Amaurornis�olivieri in the Mandrozo Protected Area,
western Madagascar, Ostrich, 89:2, 109-115, DOI: 10.2989/00306525.2017.1317296
To link to this article:
Published online: 28 May 2017.
Submit your article to this journal
Article views: 51
View related articles
View Crossmark data
Ostrich 2018, 89(2): 109–115
Printed in South Africa — All rights reserved
Copyright © NISC (Pty) Ltd
ISSN 0030–6525 EISSN 1727-947X
Ostrich is co-published by NISC (Pty) Ltd and Informa UK Limited (trading as Taylor & Francis Group)
Nesting biology and food habits of the endangered Sakalava Rail
Amaurornis olivieri in the Mandrozo Protected Area, western Madagascar§
Yverlin Z Pruvot1*, Lily-Arison René de Roland1, Gilbert Razafimanjato1, Marius PH Rakotondratsima1, Aristide
Andrianarimisa2 and Russell Thorstrom3
1 The Peregrine Fund Madagascar Project, Antananarivo, Madagascar
2 Department of Zoology and Biodiversity, Faculty of Science, University of Antananarivo, Antananarivo, Madagascar
3 The Peregrine Fund, Boise, Idaho, USA
* Corresponding author, email:
We studied the nesting biology and food habits of the endangered and endemic Sakalava Rail Amaurornis olivieri
from July to November 2015 in the Mandrozo Protected Area, western Madagascar. Three nesting pairs were
observed and their nests were constructed in a dense mat of reeds Phragmites mauritianus and averaged 56.7 ±
15.3 cm above the water (n = 3 nests). Nests were built by both adults and it took 3 d on average to complete a nest
(n = 2 nests). Thirteen matings were observed and lasted 4.1 s on average (n = 2 pairs). Average clutch size was
three eggs (n = 2 nests). Both sexes incubated; the incubation period was 15–17 d (n = 2 nests). Both male and
female participated in brooding and feeding the young, which remained for 3 d in the nest and became independent
of their parents at 45 d of age. Based on 194 identified food items, the Sakalava Rail’s diet was composed of
invertebrates: spiders (53.1%), insects (32%), crustaceans (10.8%) and molluscs (4.1%). The home ranges of two
radio-tagged individuals were 0.95 and 1.98 ha.
Nidification et régime alimentaire du Râle d’Olivier Amaurornis olivieri dans l’Aire Protégée
Mandrozo, partie ouest de Madagascar
L’étude de la nidification et du régime alimentaire du Râle d’Olivier Amaurornis olivieri, une espèce en danger
et endémique de Madagascar, a été effectuée du mois de juillet au mois de novembre 2015 dans l’Aire Protégée
Mandrozo, partie Ouest de Madagascar. Trois couples ont été observés dans leurs nids qui sont placés dans des
touffes de roseaux Phragmites mauritianus et se situent en moyenne à 56,7 ± 15.3 cm (n = 3 nids) au-dessus de
la surface de l’eau. La construction du nid qui a duré en moyenne trois jours (n = 2 nids) est assurée par les deux
sexes. Treize accouplements ont été observés et durent en moyenne 4,1 secondes (n = 2 couples). La taille de ponte
est de trois œufs par nichée (n = 2 nids). Les deux sexes ont participé à l’incubation des œufs durant 15 à 17 jours
(n = 2 nids). Le mâle et la femelle sont impliqués dans la couvaison et nourrissage des jeunes. Les poussins restent
trois jours dans le nid et sont complètement indépendants de leurs parents à l’âge de 45 jours. Basé sur 194 nourri-
tures identifiées, le Râle d’Olivier se nourrit essentiellement des invertébrés : Araignées (53,1 %), Insectes (32 %),
Crustacés (10,8 %) et Mollusques (4,1 %). Les surfaces des domaines vitaux des deux individus radiopistés sont
respectivement 0,95 ha et 1,98 ha.
Keywords: endemic, food habits, Mandrozo Protected Area, nesting biology, Sakalava Rail
The Sakalava Rail Amaurornis olivieri belongs to the
Rallidae family and is an endangered and endemic
waterbird of Madagascar (IUCN 2016). It has a small,
severely fragmented population, a restricted distribution,
and is an iconic species of wetlands in western
Madagascar. The most recent censuses were carried out
in western Madagascar wetlands between the Betsiboka
and Mangoky rivers from 2003 to 2006 at five sites:
Kinkony Lake, Ampandra Lake, Amparihy Lake, Sahapy
Lake and Mandrozo Lake, where the collective population
estimate was 215 individuals (Rabenandrasana et al.
2009). However, the population size of this species has
been estimated to be 250–999 mature individuals (BirdLife
International 2016). The restricted distribution of the
Sakalava Rail makes it vulnerable to all forms of anthropo-
genic pressure. This bird is suspected to have disappeared
from many marshes where it was once abundant and
habitat loss from conversion of wetlands to rice fields by
draining, tilling and burning, degradation of reed habitat,
the impact of introduced fish, turbidity from erosion and
§ This article is based on a paper presented at the 14th Pan-African Ornithological Conference, Dakar, Senegal, 16–21 October 2016
Published online 28 May 2017
Pruvot, René de Roland, Razafimanjato, Rakotondratsima, Andrianarimisa and Thorstrom
disturbance by fishermen and local people are all major
threats for this species (Rabenandrasana et al. 2009).
The Sakalava Rail is the least known rail species in
Madagascar. Some information about its distribution,
habitat and status have been recorded (Rabenandrasana et
al 2009), but the biology and ecology of this species are still
poorly known. Our aim was to identify the breeding season
and to describe nests and habitat requirements of this
inconspicuous rail. This study summarises new information
on the Sakalava Rail’s nesting biology and habitat, produc-
tivity, food habits and ranging behaviour, which are all
important for this species’ conservation.
Materials and methods
Study area
This study was conducted at Mandrozo Lake, in the
Mandrozo Protected Area (MPA; 17°31–17°33′ S, 44°02′–
44°06′ E), which is located in western Madagascar, in the
District of Mantirano (Figure 1). The MPA has an area of
15 145 ha and includes five ecosystems: lakes, marshes
and swamps, dry deciduous and palm forests, and savannas
(The Peregrine Fund 2013). The MPA was created by The
Peregrine Fund’s (TPF) Madagascar Project in 2007. It was
designated a Ramsar site in 2012 and it received definitive
status as an IUCN Protected Area Category V (protected
landscape) in April 2015 (The Peregrine Fund 2013). There
are two distinct seasons in MPA: a dry season from April
to October and a rainy season from November to March
(Rajaonarison et al. 2009). In 2015, the annual rainfall
and the mean average temperature recorded at the TPF
research station in Andapabe-Andranovao, at the study site,
were 1 526 mm and 27 °C, respectively.
Surveys were made at Mandrozo Lake, the fifth largest
lake in Madagascar (Figure 1). This lake has an area of
1 800 ha. It is a permanent lake of shallow fresh water and
is surrounded by reed vegetation, which provides habitat for
several bird species.
Data collection
This study was conducted during one breeding season,
from July to November 2015. We searched for individuals
and nesting pairs of Sakalava Rails in the reed vegeta-
tion surrounding Mandrozo Lake by walking slowly along
the lakeshore or using a small fibreglass canoe to enter
the openings of the reed marshes. Every 100 m travelled,
Nest 3
44°10' E44°5' E44°0' E
17°30' S
17°35' S
17°35' S
PA boundary
Nest 1
Nest 2
5 km
Figure 1: Map of three Sakalava Rail Amaurornis olivieri nests at Mandrozo Lake, western Madagascar. PA = protected area
Ostrich 2018, 89(2): 109–115
we stopped for 10 min to observe and listen for Sakalava
Rail calls (Pruvot 2016). We recorded all individuals
found or heard. When we did not observe or hear any
individuals, we used playback of the species’ typical call
using a Sony TCM 500DV tape recorder (Lor and Malecki
2002; Rabenandrasana et al. 2009). We also showed local
people, especially fishermen, a photograph of a Sakalava
Rail and asked them where they had observed this species.
Nests were discovered by tracking individuals of a located
pair until they were at their suspected nesting territory. We
also searched all habitat suspected to hold a nest.
After a nest was located, we started observations using
binoculars (10 × 42) at a distance of 6–8 m away from the
nest. We carried out daily nest observations from 06:00 to
18:00 from courtship to dispersal of young. We recorded
behaviour and activities during the nesting season:
courtship, mating, nest construction, egg laying, incubation,
hatching and rearing of young. We took egg measurements
(length and width) using vernier calipers and mass with a
Pesola spring balance scale.
The Sakalava Rail diet was identified based on direct
observation and photographs taken during the incubation
period and observations of food delivered by adults to young.
Sakalava Rail adults were captured with a net (3 m wide
by 1 m high with a 30 mm mesh) placed near the nest site.
Captured individuals were weighed, measured and ringed.
Two individuals from two different nest sites were radio-
tagged with a VHF radio transmitter mounted on the back
of each as a backpack (Kenward 1987), to estimate their
home ranges. They were radio-tracked from August to
October 2015. During the radio-tracking, locational fixes of
each individual were recorded using a handheld Etrex 10
Garmin GPS. The home range was estimated by calculating
the minimum convex polygon (MPC) of the locations (Morh
1947; Harris et al. 1990) using ArcGIS 10.0 software (ESRI,
Redlands, CA, USA).
Nest measurements were taken after dispersal of young
to avoid disturbing the nesting birds. Nest parameters
measured were shape, height above water level, nest
height, nest width, nest internal depth and nest wall
thickness. Nest site characteristics included plants (height,
diameter, scientific name and diameter of the supporting
branches), and inter-nest distances were also recorded.
Data analysis
The results of copulation duration, egg dimension, nest
measurements were calculated as mean values and
standard deviations from the measurements taken. A
chi-square test (Fowler and Cohen 1985; Johnson 1992)
was used to compare the quantity of food delivered to
chicks by both sexes. The amounts of time spent by males
and females in incubation were compared using a Student’s
t-test. Statistical analysis was performed with STATISTICA
6.0 (StatSoft, Tulsa, OK, USA) and XLSTAT 2015.1
(Addinsoft, New York, USA).
During the study period we located three breeding pairs
identified as P1, P2 and P3 with their respective nests (N1,
N2 and N3). They were discovered respectively on 14 July,
4 September and 22 September 2015. Nest observations
totaled 1 008 h: 72 h during pair formation and courtship,
48 h during nest building, 384 h during incubation and 504 h
during the young rearing period (from hatching to dispersal
of young).
Pair formation and courtship period
From mid-August to early September, we followed only
the pair P2 and observed the following behaviour. The
pair was found together at one site: feeding, allopreening,
vocalising and carrying nesting material. During this early
courtship period, the male led and presented to the female
several possible nest sites. The pair moved from one place
to another until the male stopped at a site or old nest and
called loudly, inviting the female to visit the site. When
the female arrived at the site, the male began to arrange
nesting material trying to entice her into the site. If the
female did not find the site suitable for a nest she would
not assist in nest building, which caused the male to move
to another site. When the female approved of the site she
immediately assisted in nest construction. On several
occasions (n = 6), the male visited an old nest but his mate
refused to go in. During the courtship period, we did not
observe the male feed the female.
We recorded 13 matings, of which four and nine were
observed at P1 and P2, respectively. The average duration
of copulations for both pairs was 4.1 ± 0.1 s (range =
3.9–4.2 s, n = 13). Ten and three of these copulations were
recorded between 06:00–10:00 and between 14:00–16:00,
respectively. The frequency of copulations ranged from two
to five per day.
Before mating, the male and the female perched on
floating plant stems at a distance of 3–5 m apart. The female
first emitted a specific repetitive call. Then, the male moved
to the female and made a loud call. They stood side-by-
side and emitted simultaneously a specific mating call:
“prrrrrrrourourrrrrourrouououruru…” followed by the female
moving around with the male pursuing her with a jerky step,
his neck stretched out and his head moving forwards and
backwards. When the female stopped, the male groomed
her neck and head feathers, ran around her and bowed in
front of her on each pass. When she crouched down the
male mounted her without pecking her head or neck. During
mating the female called loudly and the male emitted a
chuckle call. After mating, both sexes stood side-by-side and
allopreened. All pairs stopped mating one day before the
first egg was laid.
Nest building
We monitored nest building activity at N1 and N2 from the
start to the completion of the nest. N3 was a completed nest
when discovered. Nest building lasted 2 d for N1 and 4 d for
N2. N1 was built on 14–15 July 2015, and N2 construction
occurred from 8 to 11 September 2015. Both sexes partici-
pated in nest construction, and of 201 nest material deliveries
observed, the males and females delivered 71% (n = 142
items) and 29% (n = 59 items), respectively. Nest building
took place between 07:00 and 09:00. All nesting material was
collected 7–20 m from the nest. Nest material was composed
of dead leaves of Phragmites mauritianus, Panicum sp.,
Typha angustifolia and Cyperus aequalis (Figure 2).
Pruvot, René de Roland, Razafimanjato, Rakotondratsima, Andrianarimisa and Thorstrom
Nest and nest site characteristics.
All nests were constructed in a dense mat of reeds
Phragmites mauritianus and were situated 56.7 ± 15.3 cm
above water level (n = 3 nests, range = 40–70 cm). Nests
were placed in tufts of reeds 1–6 m in height and with stems
1–3 cm in diameter. They were supported by 10–18 reed
stems (n = 3 nests). Nest measurements and characteristics
are summarised in Table 1. The three nests were located at
a distance of 68.7 ± 32.0 m (n = 3 nests, range = 36–100 m)
from the lakeshore. The average distance between two
neighbouring nests was 1 331 ± 965.9 m (n = 3 nests, range
648–2 014 m).
The three nest sites were relatively far from the nearest
village and fishermen camps. The distance varied from
1 500 to 2 300 m. However, on several occasions we
observed fishermen near the nest sites, which disturbed the
birds and sometimes interrupted their nesting activities.
Egg laying
Egg laying occurred from mid-July to September. The
earliest recorded laying date was 16 July 2015 and the
latest was 25 September 2015. The clutch size was three
eggs in N1 and N3, and no eggs were laid in N2 despite
P2 being observed mating. All eggs were laid at one-day
intervals at both N1 and N3. The three eggs in N1 were
laid on 16, 17 and 18 July 2015 and at N3 on 23, 24 and
25 September 2015.
The mean dimensions of the three eggs at N1 were
31.4 ± 0.15 mm × 26 ± 0.14 mm and their average mass
was 14.8 ± 0.21 g (Table 2). Eggs were pale whitish with
brown spots (Figure 3).
Of the 384 h of observation during the incubation period,
180 h were at N3 and 204 h at N1. Incubation started on
completion of the clutch, and both males and females
incubated. At N1, the female incubated for 51.5%
(n = 105 h), the male for 43.1% (n = 88 h) and the nest
was unattended for 5.4% (n = 11 h) of the observation
time (n = 204 h). At N3, the female incubated for 53.3%
(n = 96 h), the male 38.8% (n = 70 h) and the pair was
absent from the nest for 7.8% (14 h) of the 180 h observa-
tion time. There was no signicant difference between
times spent by males and females in incubation at the two
nests (Student’s t-test: t = −1.10, p = 0.28). The incubation
period was 15 d at N3 and 17 d at N1. During the incuba-
tion period, the non-incubating bird did not supply food to
the incubating bird. Incubation exchanges allowed the
non-incubating individual to rest and feed.
The eggs hatched on 2 August 2015 at N1 and 8 October
2015 at N3. All eggs hatched on the same day within a one-
hour window for both nests. Unfortunately, all hatchlings at
N1 died 4 h after hatching due to an invasion and attack
of ants, which gained access to the nest by travelling on
dead and bent reed stems. Afterwards, P1 abandoned this
nest. We documented P1 renesting on 9 September 2015
with three young about 7 d old, 38 d after their first nesting
attempt failed. This confirms that Sakalava Rails can renest
after failing in the first nesting attempt.
Development of young
The development of young was only observed at the P3
nest. Both adults were involved in brooding and feeding
the young. Adults brooded young most on the hatching
day. The young spent 3 d in the nest and during 19.4
h of brooding observation, the female brooded for 73%
(n = 14.2 h) and the male for 27% (n = 5.2 h). The adults
began to feed their young on the day after hatching. We
recorded 161 food items delivered by adults to young. The
male delivered 53% (n = 85 food items) and the female 47%
(n = 76 food items). There was no significant difference
between the amount of food delivered by both sexes
during the feeding of young (chi-square test: χ2 cal = 0.008
< χ2 tab = 3.84, df = 1, p = 0.93). At age 4 d, the young
left the nest, followed their parents during the daytime and
returned to the nest to roost with one of the adults during
the night. During the first week after hatching, the young
were fed directly by adults. At age 10 d the young began to
Figure 2: A nest of Sakalava Rail Amaurornis olivieri at Mandrozo
Lake, western Madagascar
Characteristics Nest 1 Nest 2 Nest 3 Mean ± SD
Length (cm) 14.5 16 13.7 14.7 ± 1.2
Width (cm) 13.75 15.60 13.33 14.2 ± 1.2
Internal depth (cm) 5.1 7 6.2 6.1 ± 1
Heigth (cm) 13 16 11 13.3 ± 2.5
Nest wall thickness (cm) 0.75 0.40 0.37 0.5 ± 0.2
Height above water level (cm) 40 60 70 56.7 ± 15.3
Distance from lakeshore (m) 100 70 36 68.7 ± 32.0
Table 1: Characteristics of three nests of Sakalava Rail Amaurornis
olivieri at Mandrozo Lake, western Madagascar from July–
November 2015
Egg 1 Egg 2 Egg 3 Mean ± SD
Mass (g) 15 14.5 14.7 14.8 ± 0.21
Length (mm) 31.5 31.2 31.4 31.4 ± 0.15
Width (mm) 26.2 26.1 25.9 26 ± 0.14
Table 2: Mass and dimensions of three eggs of Sakalava
Rail Amaurornis olivieri in nest N1 at Mandrozo Lake, western
Ostrich 2018, 89(2): 109–115
feed themselves, but the adults still provided food until they
were 40 d old. From 41 d, the adults stopped feeding the
young and started to chase them away. At 45 d of age, the
young became completely independent and left their natal
nesting area (n = 3 young). Only the adults were observed
tending the young and no helpers or extra birds were
observed during the nesting period at these two active nest
sites. Thus, during the study period, we found no evidence
of cooperative breeding or helping at the nest by birds from
a previous brood.
Reproductive success
In the two productive nesting attempts (N1 and N3), six
eggs were laid (three eggs per nest) and all hatched,
while no eggs were laid in N2. All three hatchlings at N1
died, whereas all three young at N3 fledged and became
independent. Reproductive success was therefore
1.5 young per productive nesting attempt and 1.0 young per
territorial pair.
Food habits
During the study, 194 prey items were identified of which
33 were recorded during the incubation period and 161
during the nestling period. Spiders comprised 53.1% (103)
of identified prey, insects 32% (62), crustaceans 10.8% (21)
and molluscs 4.1% (08).
The home ranges of two radio-tagged individuals (the
female of P2 and the male of P3) were respectively 0.95 ha
(n = 39 locational fixes) and 1.98 ha (n = 60 locational
fixes). There was no overlap between the territories of these
two radio-tagged birds.
Despite the huge effort in searching for pairs and nesting
activity only three pairs were located and studied at MPA.
Several reasons may explain why we found so few pairs:
the population size of Sakalava Rails at Mandrozo Lake is
extremely low, Sakalava Rails are secretive and reclusive
species making detection difficult and the habitat is
characterised by dense and sometimes impenetrable reed
vegetation that makes observation difficult. In addition,
there is a possibility that this species has a longer breeding
season, maybe year round, and we may have missed part
of the nesting period because we only made observations
from July to November. Furthermore, there is evidence
that Sakalava Rails breed in the wet season with young
observed in February and March and egg laying in March
(Rand 1936; Benson and Wagstaffe 1972). These records,
and our study during the dry season, confirm that Sakalava
Rails have a protracted breeding season in both the wet
and the dry seasons, similar to the Black Crake Amaurornis
flavirostris (Taylor and van Perlo 1998).
The events that occur before and at the beginning of pair
formation are often difficult to observe in a bird’s reproduc-
tion period (Gill 1990), and our observations during this
study were consistent with this statement. We observed
pair formation between August and September, whereas
Rabenandrasana (2007) observed it in November at
Amparihy Lake, 100 km north of MPA. However, we found
that pair formation occurred from July to September. If
we consider Rabenandrasana (2007) and this study, we
conclude that pair formation of Sakalava Rails takes place
from July to November. During pair formation, both sexes
accompanied each other frequently to seek a suitable
place for nesting. This behaviour has also been observed
in several species of rail, including the Black Crake (Taylor
and van Perlo 1998).
The three nests described during this study were newly
built even though we did not see the construction of N3.
The fresh state of the N3 nest material suggested it was
a new nest. Rabenandrasana et al. (2009) observed that
Sakalava Rails built a new nest every breeding season.
Taylor and van Perlo (1998) reported that most rail species
do not reuse an old nest. We observed that both sexes
of the Sakalava Rail participated in nest construction,
and this is extremely common among the Rallidae family
(Bouglouan 2010). The average duration of nest construc-
tion was 3 d during this study. For other rails, it varies
from 3 to 5 d for Moorhens Gallinula chloropus (Hemery
and Blaize 2013) and 4 d for Water Rails Rallus aquaticus
(Guillemont and Koenig 1994). Nest construction duration
varies among species and depends on nest characteris-
tics and habitat (Silva et al. 2010). Benson and Wagstaffe
(1972) found a nest of Sakalava Rail on Typha angustifolia
50 cm above ground level. Another nest was discovered
in a deep tunnel of Phragmites 70 cm above the ground
(Rabenandrasana 2007). In this study, we observed three
nests placed in Phragmites mauritianus and averaging
56.7 cm above water level. We believe that Phragmites
is predominantly used by Sakalava Rails for building their
nests. The average distance between neighbouring nests
of Sakalava Rails at MPA was 1 331 m, which is much
greater compared with other species of rail such as Water
Rail (20–50 m) (Taylor and van Perlo 1998) and Virginia
Rail Rallus limicola (46 m) (Conway 1995). This greater
distance may reflect a low population size and density of
Sakalava Rail at Mandrozo Lake while suitable habitat still
exists and is available. Bouglouan (2010) reported that
most rails nest solitarily and well separated, except when
nesting habitat is scarce.
Figure 3: Nest of Sakalava Rail Amaurornis olivieri with a
three-egg clutch at Mandrozo Lake, western Madagascar
Pruvot, René de Roland, Razafimanjato, Rakotondratsima, Andrianarimisa and Thorstrom
Despite the fact that the nest sites were far from human
habitation, they were disturbed by fishermen who came
frequently to fish near the reeds bordering the lake. This
disturbance, especially the noise they made driving fish by
whacking their canoes, could cause nest abandonment by a
pair as reported by Rabenandrasana et al. (2009).
The home ranges of the two radio-tagged individuals were
0.95 ha and 1.98 ha. These home ranges may be small
because they were recorded during the incubation period.
Jedlikowski and Brambilla (2017) state that small home
ranges are due to variations in incubation behaviour (some
individuals spend more time incubating clutches and have,
on average, smaller home ranges) and food availability at the
site. However, this result seems to be normal compared with
other species of rails: Sora Porzana carolina (0.20–0.81 ha)
(Johnson and Dinsmore 1985; Conway 1990), Water Rails
(671.6 m2 on average) and Little Crakes P. parva (447.3 m2
on average) (Jedlikowski and Brambilla 2017).
Rabenandrasana et al. (2009) reported that the peak
laying period for Sakalava Rail was between July and
November, and we observed laying from July to September
at Mandrozo Lake. However, Benson and Wagstaffe
(1972) found a nest of this species with two eggs in March
at Kinkony Lake, 240 km north-east of MPA. Robertson
(2004) announced the possibility that the egg-laying period
extends until the wet season. Further study is needed to
verify this long breeding season and to establish whether
the species breeds all year round. Regarding clutch size,
we recorded three eggs in each of two nesting attempts.
Rabenandrasana et al. (2009) found two eggs at Mandrozo
Lake in October 2005 and a nest with three eggs was
located at Amparihy Lake. The modal clutch size of
three eggs for the Sakalava Rail is small compared with
other species in the same genus such as White-breasted
Waterhen Amaurornis phoenicurus (4–9 eggs per nest)
(Le-Dantec 2015), Pale-vented Bush-hen Amaurornis
moluccana with a clutch size from four to seven eggs
(Le-Dantec 2009), and Brown Crake Amaurornis akool
(5–6 eggs) (Le-Dantec 2014). This difference may be
due to the variation in size between the species. Taylor
and van Perlo (1998) confirmed that clutch size increases
with the physical size of the species for many rails. The
aforementioned species, having respectively a size of
33 cm, 30 cm and 28 cm, are larger than Sakalava Rail
(19 cm on average).
Incubation started on completion of the clutch, as occurs
in many rails (Taylor and van Perlo 1998). In this study the
incubation period of the Sakalava Rail was 15–17 d, similar
to that of the Black Crake (13–19 d; Taylor and van Perlo
1998). Bouglouan (2010) mentioned that incubation usually
lasts from 15 to 19 d in the Rallidae.
With regard to the death of the N1 chicks due to an
attack by ants, it is difficult to say if this is a common cause
of nest failure because it was only observed in one nesting
attempt, but it seems likely to be a rare event as the rails
nest over water.
During this study, both sexes participated in brooding and
feeding of young, which is common nesting behaviour in
the Rallidae (Bouglouan 2010). We believe that coopera-
tive breeding or helping at the nest by birds from a previous
brood does not occur in the Sakalava Rail as the young
were being chased away by their parents causing them to
disperse from their natal nesting area at 45 d of age.
After hatching, the young precocial Sakalava Rails spent
3 d in the nest before leaving it, which is common for
most rails (Taylor and van Perlo 1998). Young Common
Moorhens Gallinula chloropus stayed in the nest for at least
2 d (Wood 1974). However, young of the Water Rail are
able to leave the nest on the day of hatching (Guillemont
and Koenig 1994). Sakalava Rail young were completely
independent of the adults and dispersed from their natal area
at 45 d of age. For other species, young disperse at 60 d of
age for Water Rails, 25 d of age for Spotted Crakes Porzana
porzana and 45–50 d for Little Crakes (Godin 2003).
Rabenandrasana and Zefania (2005) reported that
the diet of Sakalava Rails is composed of crustaceans
and invertebrates. During this study, Sakalava Rails
consumed spiders, insects, crustaceans and molluscs.
Spiders made up the greatest portion of their diet at
53.1%. However, Black Crakes feed on a very wide range
of aquatic vertebrates and invertebrates, including worms,
crustaceans, molluscs, insects and their larvae, small fish,
small frogs and tadpoles (Taylor and van Perlo 1998). They
may also take eggs and nestlings of other birds species
(herons and weavers), they consume seeds and parts of
aquatic plants, and they scavenge carcasses of crabs,
crayfish and small birds (Taylor and van Perlo 1998). We
believe that Sakalava Rails also eat other food types, but
we did not record any during our observations. It would be
good to conduct a specific study focused only on their diet
and using camera traps and whitewash analysis in order to
collect new information on their diet.
Despite only three pairs of Sakalava Rails being observed
in this study, new information was obtained on several
aspects of the nesting biology and breeding season of
this threatened species. We found that both males and
females took part in all nesting activities during the various
stages of the breeding period: from nest building to rearing
of young. In addition, their breeding period coincided with
the greatest abundance of invertebrates that facilitated
the feeding of young (ZYP pers. obs.). The Sakalava Rail
requires Phragmites reeds and other dense aquatic vegeta-
tion for nesting and other activities. This study was only
a short-term preliminary study of the Sakalava Rail, and
we recommend further studies to determine the species’
population size, status and ranging behaviour during the
post-breeding and non-breeding periods, and to elucidate
more on the species’ reproductive strategy and the
mechanisms that regulate its population.
Acknowledgements — This study was supported by The Peregrine
Fund’s Madagascar Project with funding provided, in part, by the
Madagascar Biodiversity Fund. Our special thanks go to these
both institutions. We would like to thank the Mandrozo Protected
Area technicians of The Peregrine Fund for their assistance in data
collection in the field. A special thanks to Barry Taylor for reviewing
and improving earlier versions of this manuscript.
Benson CW, Wagstaffe R. 1972. Porzana olivieri and Limnocorax
flavirostris: a likely afnity. Bulletin of the British Ornithologists’
Club 92: 160–164.
Ostrich 2018, 89(2): 109–115
BirdLife International. 2016. Species factsheet: Zapornia olivieri.
Available at [accessed 8 December 2016].
Bouglouan N. 2010. Les Rallidés. Available at http://www.oiseaux- [accessed 24 November 2015].
Conway CJ. 1990. Seasonal changes in movements and habitat
use in three sympatric rails. MSc thesis, University of Wyoming,
Laramie, USA.
Conway CJ. 1995. Virginia Rail Rallus limicola. In: Poole A, Gill
F (eds), The Birds of North America no. 173. Philadelphia:
Academy of Natural Sciences; Washington, DC: American
Ornithologists’ Union. pp 1–20.
Fowler RA, Cohen L. 1985. Statistics for ornithologists. London:
British Trust for Ornithology.
Gill FB. 1990. Ornithology. New York: WH Freeman and Company.
Godin J. 2003. Les oiseaux des roselières du bassin Artois-
Picardie. Partez à la rencontre de la Biodiversité. Douai: Agence
de l’eau Artois- Picardie.
Guillemont A, Koenig P. 1994. Le Râle d’eau Rallus aquaticus. In:
Yeatman-Berthelot D, Jarry G (eds), Nouvel atlas des oiseaux
nicheurs de France 1985–1989. Paris: Société Ornithologique de
France. pp 242243.
Harris RB, Cresswell WJ, Forde PG, Trewhella WJ, Woollard T,
Wray S. 1990. Home-range analysis using radio-tracking data –
a review of problems and techniques particularly as applied to
the study of mammals. Mammal Review 20: 97123.
Hemery D, Blaize C. 2013. Suivi des rallidés sur les pièces d’eau
du Bois de Vincennes et des lacs périphériques : foulque
macroule Fulicaatra et Gallinule poule d’eau Gallinula chloropus.
Rapport d’activité. Pont-de-Buis-lès-Quimerch: Grumpy Nature;
Paris: Mairie de Paris.
IUCN (International Union for the Conservation of Nature and
Natural Resources). 2016. The IUCN Red List of Threatened
Species. Version 3.1. Available at,
[accessed 8 December 2016].
Jedlikowski J, Brambilla M. 2017. Effect of individual incubation
effort on home range size in two rallid species (Aves: Rallidae).
Journal of Ornithology 158: 327332.
Johnson RR. 1992. Elementary statistics (6th edn). Boston, MA:
Johnson RR, Dinsmore JJ. 1985. Brood-rearing and postbreeding
habitat use by Virginia rails and soras. Wilson Bulletin 97:
Kenward R. 1987. Wildlife radio tagging. London: Academic Press.
Le-Dantec D. 2009. Râle des Moluques Amaurornis moluccana.
Available at [accessed 12 December
Le-Dantec D. 2014. Râle akool Amaurornis akool. Available at[ accessed 14 December 2015].
Le-Dantec D. 2015. Râle à poitrine blanche Amaurornis
phoenicurus. Available at [accessed
16 December 2015].
Lor S, Malecki RA. 2002. Call-response surveys to monitor marsh
bird population trends. Wildlife Society Bulletin 30: 11951201.
Morh CO. 1947. Table of equivalent populations of North American
small mammals. American Midland Naturalist 37: 223249.
Pruvot YZM. 2016. Bio-écologie et abondance du Râle d’Olivier
Amaurornis olivieri (Grandidier & Berlioz, 1929) dans l’Aire
Protégée Mandrozo, district de Maintirano, région Melaky.
Mémoire de DEA en Biodiversité et Environnement, Université
de Toliara, Madagascar.
Rabenandrasana M. 2007. Conservation biology of Sakalava rail
Amaurornis olivieri an endangered Malagasy water bird and
public awareness in Besalampy wetlands complex, western
Madagascar. Final report. Antananarivo: ASITY (Ligue Malgache
pour la protection des Oiseaux).
Rabenandrasana M, Zefania S. 2005. Population size, distribution,
habitat use and threats of Sakalava rail Amaurornis olivieri
surveyed between Mahajanga and Toliary, Western Madagascar
(August 2003April 2004). Final report. Antananarivo:
Conservation International Madagascar and ASITY (Malagasy
Bird Club).
Rabenandrasana M, Zefania S, Long P, The Seing S, Virginie MC,
Randrianarisoa M, Safford R, Székely T. 2009. Distribution,
habitat and status of 'Endangered' Sakalava Rail of Madagascar.
Bird Conservation International 19: 2332.
Rajaonarison JL, Razanameharizaka JH, Randrianirina EF. 2009.
Etude d’impact environnemental et social de la future Nouvelle
Aire Protégée de Mandrozo, District de Maintirano. Antananarivo:
The Peregrine Fund.
Rand AL. 1936. The distribution and habits of Madagascar birds.
Summary of the field notes of the Mission Franco-Anglo-
Américaine à Madagascar. Bulletin of the American Museum of
Natural History 72: 143499.
Robertson I. 2004. First-ever photographs of Sakalava Rail
Amaurornis olivieri and first detailed observations since 1962.
Bulletin of the African Bird Club 11: 1821.
Silva B, Correia J, Nunes F, Tavares P, Varum H, Pinto J. 2010.
Bird nest construction - lessons for building with earth. WSEAS
Transactions on Environment and Development 6: 8392.
Taylor B, van Perlo B. 1998. Rails: a guide to the rails, crakes,
gallinules and coots of the world. Robertsbridge: Pica Press.
The Peregrine Fund. 2013. Plan d’Aménagement et de Gestion
de la Nouvelle Aire Protégée de Mandrozo. Antananarivo: The
Peregrine Fund, Madagascar.
Wood NA. 1974. The breeding behaviour and biology of the
Moorhen. British Birds 67: 104115.
Received 12 January 2017, revised 24 March 2017, accepted 3 April 2017
Associate Editor: Mark Brown
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Home range varies with individual traits, environment, and time, but studies relating variations to behavioural traits are scarce. We investigated variation in home-range size in nesting rallids in relation to incubation effort, sex, weight, available space, and food abundance. In Water Rail, individuals more involved in incubation had smaller home-ranges, regardless of sex, whereas in Little Crake both parents incubate with similar effort, and the main factor positively affecting home-range size was available space. Results suggested that home-range size during the nesting period may vary according to incubation behaviour, especially in species with high inter-individual variation in incubation effort.
Habitat use by breeding Virginia rails (Rallus limicola) and soras (Porzana carolina) in northwestern Iowa was correlated highly with the availability of emergent vegetation for both species. Virginia rails preferred fine and moderately robust emergents, and soras were most abundant at diverse shoreward sites. Monotypic stands of glaucous cattail (Typha glauca) received the greatest percentage of use by both species. Structurally, habitat was used as available. No strong niche-segregating mechanism was evident in habitat use. Practices used to encourage waterfowl use are compatible with habitat requirements of breeding Virginia rails and soras. Drawdowns may be used to create attractive emergent diversity and appropriate cover-water interspersion.
The volume is broadly split into two main sections. The firsts consists of seven introductory chapters: biodiversity and priority setting; identifying endemic bird areas; global analyses; the prioritization of endemic brid areas; the conservation relevance of endemic bird areas; endemic bird areas as targets for conservation action; and regional introductions. The second, and larger part of the text looks at the endemic bird areas in detail. The section is split into six subsections, by region: North and Central America; Africa, Europe and the Middle East; continental Asia; SE Asian Islands, New Guinea and Australia; and the Pacific Islands. Within each regional subsection the endemic areas are detailed, providing information on : general characteristics; restricted-range species; threats and conservation; and location maps.
Most marsh birds are inconspicuous, occupy dense habitats, and are difficult to detect by sight or vocalization. They are seldom recorded by local or national monitoring efforts, such as the United State Geological Survey's Breeding Bird Survey. Taped call-response surveys can increase the detection rates of secretive birds such as American bitterns (Botaurus lentiginosus), pied-billed grebes (Podilymbus podiceps), soras (Porzana carolina), and Virginia rails (Rallus limicola). Using tape-playback calls, we demonstrated that the sample size required to effectively establish a monitoring program to detect a 5% population change in 5 secretive marsh-bird species over a 5-year period was 24 stations for American bitterns, 28 for least bitterns, 31 for pied-billed grebes, 27 for soras, and 13 for Virginia rails.
The ‘Endangered’ Sakalava Rail Amaurornis olivieri is endemic to wetlands in western Madagascar, where it has been recorded between the Betsiboka river in the north and the Mangoky river in the south. Between August 2003 and November 2006, including dry and wet seasons, we surveyed 36 potentially suitable wetlands throughout its known range. We found Sakalava Rails at five sites: Lacs Kinkony, Ampandra, Amparihy, Sahapy and Mandrozo. At each site the population was small (12–39 individuals) and the highest density was 20 individuals km−2. We found up to 67 birds in each field visit and the total number of birds (sum of maxima at each site) seen was 100. We estimate the total population at the five sites to be 215 rails. We cannot confirm that the population lies within the range estimated in the current Red List (250–999 individuals), although this may yet be proven correct. The typical breeding habitat of Sakalava Rail is lotic marshes with a mixture of large areas of open water, reed Phragmites mauritianus and floating Salvinia hastata. The major threats to Sakalava Rail appear to be habitat loss caused by wetland conversion to rice fields and by fires, disturbance by fishermen and people from local villages, and hunting. Other processes that may alter the ecological character of wetlands and so affect their suitability for Sakalava Rails, such as hydrological change or the effects of exotic fish or vegetation, remain to be investigated.
The Madagascar plover Charadrius thoracicus is a shorebird endemic to western Madagascar, currently classified as globally vulnerable. It is restricted to specialized wetland habitats that are increasingly threatened by humans. To inform future conservation measures for this poorly known species, we develop a predictive habitat suitability map and use this map to estimate the size of the Madagascar plover population. We integrate spatially referenced presence-only observations of Madagascar plovers with Landsat data, elevation data and measures of distance to settlements and the coast to produce a habitat suitability model using ecological niche factor analysis. Validation of this model using a receiver operating characteristic plot suggests that it is at least 84% accurate in predicting suitable sites. We then use our estimate of total area of suitable habitat above a critical suitability threshold and data on Madagascar plover density in suitable sites to estimate the total population size to derive a total population estimate of 3100±396 standard error individuals. Finally, we explore the conservation applications of our model.