See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/288095684

Roosting and nest-building Behaviour of the

white-nest Swiftlet Aerodramus Fuciphagus
(Thunberg) (Aves: Apodidae) in farm....

Article in The Raffles Bulletin of Zoology · November 2013

CITATIONS READS

3 370

3 authors, including:

Mohamad Fizl Sidq Ramji Mustafa Abdul Rahman

University Malaysia Sarawak University College Sabah Foundation
10 PUBLICATIONS 21 CITATIONS 50 PUBLICATIONS 329 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Genomics of Bornean diversity: sequencing and whole genome population analysis of critically
endangered species View project

All content following this page was uploaded by Mohamad Fizl Sidq Ramji on 27 April 2016.

The user has requested enhancement of the downloaded file.

 THE RAFFLES BULLETIN OF ZOOLOGY 2013

THE RAFFLES BULLETIN OF ZOOLOGY 2013 Supplement No. 29: 225–235

http://zoobank.org/urn:lsid:zoobank.org:pub:97E71EFC-291F-4DA5-98D4-C85466A3E362
Date of Publication: 30 Nov.2013
© National University of Singapore

ROOSTING AND NEST-BUILDING BEHAVIOUR OF THE WHITE-NEST SWIFTLET

AERODRAMUS FUCIPHAGUS (THUNBERG) (AVES: APODIDAE)
IN FARMED COLONIES

Mohamad Fizl Sidq Ramji

Department of Zoology, Faculty of Resource Science and Technology, Universiti Malaysia Sarawak
94300 Kota Samarahan, Sarawak
Email: f_sidq@yahoo.com (Corresponding author)

Lim Chan Koon

157 Lorong 4A, Off Jalan Stampin Timur, 93350 Kuching
Email: limchankoon@yahoo.com

Mustafa Abdul Rahman

Research and Innovation Management Centre, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak
Email: rmustafa@rimc.unimas.my

ABSTRACT. — The edible-nest swiftlets of the genus Aerodramus are amongst the most unusual of birds,
being able to navigate in total darkness aided by echolocation and using their own saliva to construct the
nest. They are a valuable economic resource, the edible nests being much sought after. Knowledge of nesting
and breeding ecology of this species has so far been limited to cave colonies whilst studies focusing on the
house-farmed population are lacking. We studied the roosting and nest building behaviour of the white-nest
swiftlet Aerodramus fuciphagus (Thunberg) in two separate house-farmed colonies of different age in Miri
Division, Sarawak, from Jun.2010 to Jan.2011 (Site-I) and Feb.2012 to Oct.2012 (Site-II). Two types of infra-
red (IR) cameras were used, namely (i) fixed focal-lens IR to monitor large colony and (ii) Pan-Tilt-Zoom
camera for close-up observation. This paper reports new discovery in which three basic activity sessions
are described; first emergence period (0600–0700 hours), post-emergence period (0700–1000 hours) and
returning period (1800–1900 hours). During the post-emergence period, approximately half of the sampled
colony was observed re-entering the swiftlet house to resume nest construction. Ten ethogram categories
were developed to describe the roosting behaviours of the white-nest swiftlets: proximity fluttering, random
roosting flight, pair switching, parallel shifting, mounting, preening, defaecating, resting, territorial display,
and nest building. Our results also revealed that there is a disparity in sexual contribution in nest building,
where one partner is twice more hardworking and return more frequently during the post-emergence period
to build nest. We hypothesized that it is the male (i.e., Individual-A) that contributes more to nest building,
reasons being (i) Individual-A is the one that mounted Individual-B and not the other way around, (ii)
Individual-A is nearly twice as hardworking in nest building, correlating with the fact that spermatogenesis
is less energy demanding than oogenesis, and (iii) more protective over its partner when their nest reaches
full size, a point of time when copulation is expected.

KEY WORDS. — Aerodramus fuciphagus, swiftlet, edible nests, roosting, nest building, colony, sexual
contribution

INTRODUCTION species are able to construct nests using salivary nest

cement secreted from a pair of sub-lingual glands (Lim &
The white-nest swiftlet Aerodramus fuciphagus (Thunberg) Cranbrook, 2002) mixed, in most species of swiftlets, with
is one of several small cavernicolous, echolocating and material such as feathers and plant fibres. The nest of the
insectivorous swifts from the Family Apodidae (Chantler white-nest swiftlet is comprised primarily of pure saliva
& Driessens, 1995). Swiftlets are arguably the most (Cranbrook & Lim, 1999). This edible nest cement is the
accomplished flyers among all bird species, spending most main ingredient of the highly-prized and renowned “bird’s
of their lives on the wing, catching and feeding on insects nest soup” which is of significant commercial and reputed
in flight (Cranbrook & Lim, 1999). When at roost, these pharmaceutical value (Ismail, 1999; Tompkins, 1999; Lim,

225
 Ramji et al.: Roosting and nesting behaviour of swiftlets

2000). Because of their valuable nests, edible-nest swiftlets complex digital equipment such as multichannel event
have a long historical affiliation with humans from the early recorders. In contrast to conventional direct observation,
exploitation of natural cave colonies to the more enterprising video images provide objective study materials and are
undertaking of swiftlet farming in recent years. highly repeatable for intermittent or rigorous analysis (Reif
& Tornberg, 2006). In addition, remote video monitoring
Making use of the propensity of wild swiftlets to nest within is highly manageable with only minimal disturbance to the
disused buildings, swiftlet farmers now build empty house nesting colony (Johnston et al., 2003).
structures to attract new nesting colonies (Lim & Cranbrook,
2002). These can be built in towns where their valuable Given the rapid developments in this field, the use of
contents are easily guarded and the swiftlets are able to forage continuous video recording has become commonly
over the surrounding countryside. Inside the dark, empty, applicable in avian behavioural studies (Reif & Tornberg,
cave-like environments, members of the house colonies can be 2006). Monitoring of nesting and roosting behaviour using
found roosting, clinging side by side in pairs on the parallel non-stop video surveillance systems and time-lapse digital
planks and beams that have been purposely set up to entice video has built up mounting interest in recent ornithological
them to nest. Swiftlet pairs have been reported to roost and research (e.g., Pechacek, 2005; Smithers et al., 2005; Pierce
build their nest at their permanent roosting site, therefore & Pobprasert, 2007). In the past, the behavioural patterns of
implying high nest-site fidelity and pair kinship in colonial white-nest swiftlets in farmed colonies were rather subject to
swiftlets (Lim, 1999). Large numbers of individuals tend to theoretical interpretation based on observations made on their
congregate and nest within close proximity of each other, natural cave-dwelling counterparts. Unlike the wild colonies,
while those roosting far away from any nest are believed to be the swiftlets living in swiftlet-houses are highly adaptive
sexually immature juveniles or non-reproductive individuals towards habitat modifications in cave-like mimicry or limited
(Medway 1962a; Lim & Cranbrook, 2002). In general, nest roosting space on parallel wooden beams installed in close
building behaviour is commonly associated with courtship proximity. The advantages of using digital video recording
display and pair formation in birds (Soler et al., 1998a). With technique confer an excellent opportunity to unravel the
regard to the white-nest swiftlet, it was believed that both potentially complex behavioural mechanisms among colonial
sexes contribute to nest building that normally takes place swiftlets. In this study, we present the nightly ethogram of
at night while at roost (Kang & Lee, 1991). the white-nest swiftlet colony at roost with emphasis on the
division of labour in nest building.
Over the years, the edible-bird nest industry has seen an
influx in international market revenues that has prompted
the establishment of more and more “swiftlet houses” in MATERIAL AND METHODS
Borneo (Mardiastuti et al, 1997; Lim & Cranbrook, 2002).
In Sarawak, the conventional management of swiftlet Study area and video recordings. — The swiftlet colonies
houses is largely manipulated by local entrepreneurs or used in this study were located in two swiftlet houses
farm managers, most of whose practices lack a scientific (4°23'39"N, 113°59'12"E) in Miri Division, Sarawak. They
basis. Present knowledge on the biology and ecology of were identified as A. fuciphagus (Thunberg) by C. K. Lim
Bornean swiftlets (Aerodramus spp.), in particular the cave and the Earl of Cranbrook based on the point that although
populations, is reasonably well documented (e.g., Medway, the rump feathers were paler than the back, these swiftlets
1962a, 1962b, 1967; Lim, 1999, 2000; Lourie & Tomkins, were distinctly different from A. germani in which the
2000; Thomassen, 2005). However, important ecological rump feathers are greyish-white with a black shaft. Further,
information such as roosting and nesting behaviour, either in white-nest swiftlet populations found in inland caves have
caves or farmed colonies, is scarce and not clearly understood. rump colouration that is the same as the back (i.e., currently
Information such as how the swiftlets interact with each other accepted as subspecies A. fuciphagus vestitus). Therefore
at roost, whether they pair for life or are they polygamous, we identified the swiftlet populations in house farms in this
and where sexually immature fledglings roost within the study as A. fuciphagus fuciphagus. Two types of infra-red
colony are not only scientifically interesting, but are very CCTV recording equipment were used in this behavioural
useful to swiftlet farmers. Therefore, systematic ethological study; (i) fixed focal-lens Infra Red (IR) camera (1/3 Sony
study using continuous digital video recording system was Super HAD CCD-NIR-6036) and (ii) Pan-Tilt-Zoom camera
used in this study to find some insights to these questions. (Sony DN-PTZ Camera High Resolution IR-corrected
aspherical power lens) with PTZ Key3 Joystick Controller.
The study of animal behaviour describes ways in which Site-I is an eight-year old colony with more than 4,000 pairs,
animals interact with their environment and the survival value while Site-II in the adjacent building is only four years old
of that behaviour (Drickamer & Vessey, 1992). Ethology with roughly 700 pairs of swiftlets. Two IR cameras were
can be defined as the systematic biological approach to the installed in each swiftlet houses, wired via video cables to
study of animal behaviour, while an ethogram describes a four-channelled digital video recorder (H.264 4CH DVR)
an inventory of the behaviour patterns performed by the connected to a 15” inch flat screen monitor. To acquire
species under investigation (Prakash et al., 1994). In the sufficient viewing coverage of the sampled colony, the fixed-
case of modern ethological studies, there has been significant lens static IR camera scope was positioned perpendicularly
enhancement of quantitative observations using recent data facing the targeted nesting compartments. The PTZ camera
collection techniques such as remote photography and highly was wall-mounted adjacent to the IR camera in site-II. All

226
 THE RAFFLES BULLETIN OF ZOOLOGY 2013

recordings were automatically stored into the DVR hard by each individual (marked-A and unmarked-B). Similarly,
disk as video files that were periodically retrieved for later total duration was calculated in minutes by screening
viewings or analysis. Site-I was studied from Jun.2010 to every quarter-hour over 24 hours. Observational notes
Jan.2011, while recording at Site-II began from Feb.2012 to were made on specific behaviours, roosting positions and
Oct.2012. In the first site, a small colony of 20–30 (out of nest building technique displayed by the pair in response
150) breeding pairs within four central compartments was to different phase of nest completion. Comparison of nest
left unharvested while monthly harvesting was conducted building duration between the two was tested via 2-sample
in the second site. At Site-II, a pair was targeted among a independent t-test. Mean daily frequency of all classified
small breeding cluster of five pairs for close-up observation. roosting behaviours of the pair was tabulated and divided
We tested the hypothesis of equal sexual contributions to into four weeks observation. All statistical tests were carried
nest building activity by marking one of the individuals in out using Microsoft Office Excel 2007 and MINITAB 13.2
the targeted pair. Markings were done on the wing tips and (MINITAB Inc. 2000).
tail feathers with fluorescent green marker (MARK HERTM
livestock marking paint) on 20 Feb.2012. The marked
individual was designated as Individual-A and its unmarked RESULTS
partner as Individual-B. After a one-month trial of video
monitoring, the nest of this targeted pair was removed on Daily patterns of emergence and return. — Overall,
31 Mar.2012 in an attempt to dissect the specific nesting 2,688 hours of observations were analysed by tallying
behaviours within a full cycle of nest building. 10,752 quarter-hourly counts from the two colonies. Three
basic activity sessions can be described, namely the first
Data analysis. — To compare the daily activity pattern and emergence period (0600–0700 hours), post-emergence period
population size at both sites, data for Site-I was pooled from (0700–1000 hours) and returning period (1800–1900 hours).
the maximum number of recording hours available from In general, the white-nest swiftlets spent between 12–17
Jun.2010 to Jan.2011. Likewise, data for Site-II ranged from hours roosting inside the swiftlet house. The earliest sign
Feb.2012 to Oct.2012. Each roosting area (within the viewing of flight movements commenced between 0600–0615 hours.
scope of the fixed focal lens IR camera) was monitored to Subsequently, the number of individuals present decreased
calculate the total number of roosting individuals and periodic abruptly as the swiftlets exited their roosting site, but roughly
intervals of nest attendance (Chazarreta et al., 2011). All half of the estimated population would return shortly after
video footage was viewed on fast-forward mode. However, 0700 hours. A unimodal curve was observed within the
comprehensive screening was done occasionally at normal period 0700–1000 hours, peaking between 0800–0900 hours.
speed to characterise miscellaneous activities that might From 1000 hours onwards, the numbers began to decline
be potentially overlooked (Pechacek, 2005). Subsequently, again and reverted to its initial empty state leaving just the
all prescribed activities were transcribed onto activity log young fledglings at the nests (Fig. 1). Later in the afternoon,
sheets according to the 24:00 hours timeline with numbers a few individuals may gradually enter the house as early
of bird counted per observation hour. Following Martin & as 1700 hours. By 1840 hours, large swarm of swiftlets
Bateson (1993) and Pechacek (2005), a simple method of would typically rush back to their roosting sites. The frantic
instantaneous sampling (fix-interval time point) was chosen to condition normally lasted for 15 to 20 minutes. Such daily
organise and screen the data systematically. For consistency outrush and return was similar from month to month for both
and comparison, recording time was fragmented and analysed sites. Based on the numbers of returning individuals during
from four quarter-hour sampling intervals (Q1, Q2, Q3 and the post-emergence period, the percentage of return for Site-I
Q4) per observation hour. Sample points were derived from and Site-II ranged from 48–87% (Dec.2010 to Jan.2011) and
each quarter-hour to infer the general roosting behaviours 29–66% (Feb. to Oct.2012) respectively (Table 1).
and their associated movements displayed by the targeted
individuals. Using the multifocal PTZ camera, we confirmed
the stated behaviours by capturing still photos of the targeted
pair. Total number of individuals, mean and standard error
(SE) for each hour were used to plot the graph for daily
emergence and returning pattern. Finally, cumulative graphs
describing the corresponding monthly pattern for both sites
were generated from the mean numbers of nest attendance
counted for every month.

For focal observation on the targeted pair, the nest building

rate was measured by calculating the duration and total
frequency of periodic intervals during deposition of salivary
layers. Time of nest building commence from the first
indicative point of the bill movements extruding saliva
until break between any deposition of layers. The close-up
Fig. 1. Cumulative graph showing average numbers of roosting
view using multifocal PTZ lens enabled clear and detailed
individuals per observation hour between Site-I (cross markers)
verification of specific time and duration of saliva deposition and Site-II (solid circles).

227
 Ramji et al.: Roosting and nesting behaviour of swiftlets

Table 1. Maximum number of returning individuals (n), estimated population size (N) and percentage of return (%) during the post-
emergence period (0700–1000 hours) for swiftlet colony in Site-I and Site-II.

Site Date Maximum number of Estimated population Percentage of post-emergence

returning individuals (n) size (N) return (%)
I Jun.2010 143 163 87
I Dec.2010 189 246 77
I Jan.2011 143 295 48
II Feb.2012 94 142 66
II Mar.2012 49 142 35
II Apr.2012 43 136 32
II May 2012 56 136 39
II Jun.2012 38 127 30
II Jul.2012 34 128 27
II Aug.2012 50 145 34
II Sep.2012 59 148 40
II Oct.2012 42 146 29

Classification and description of ethogram. — We developed Proximity fluttering. — Fluttering in proximity to another
ten terminologies to describe the roosting behaviours of individual is among the common roosting behavioural
the white-nest swiftlets (Fig. 2). Each behaviour pattern is displays by both single and paired individuals. Within each
mutually exclusive. The mean daily frequency ± SE of the nesting compartment, swiftlet at roost will either move by
listed behaviours is presented in Table 2 with corresponding fluttering at one point or circling for a short while around their
diagrams illustrating each movement (Fig. 3). Detail nest before returning to the exact roosting area. In general,
descriptions are as follows: proximity fluttering and random roosting flights activities

Fig. 2. An ethogram chart of the white-nest swiftlet (A. fuciphagus) showing roosting behaviour classifications with their associated
movements.

228
 THE RAFFLES BULLETIN OF ZOOLOGY 2013

typically build up during the final hours (0500–0600 hours)

Table 2. Mean daily frequency performed by both Individual-A and Individual-B for proximity fluttering (PF), random roosting flights (RRF), pair switching (PSw), parallel shifting (PSh), mounting
before the first emergence from the swiftlet house. Of the

24.6±1.2
observed pair, Individual-A and Individual-B displayed this

B
1.6±0.6
1.7±0.6
2.5±0.4
6.4±0.7

3.8±0.5
0.8±0.2
Daily average
behaviour 3.4 ± 0.5 and 1.6 ± 0.6 times nightly respectively.

0
Random roosting flight. — This behaviour consists of a
continuous series of proximity fluttering or random directional

24.7±1.2
A
3.4±0.5
2.5±0.6
7.5±0.9
4.3±0.4
0.4±0.2

3.3±0.4
2.1±0.3
flights with one to four stops. Intense random flights usually
occurred during the first hour after returning to the colony
(1830–1930 hours) and gradually decreased throughout the
night. Just before dawn, the activities increased and during
this frantic period, a nesting individual may occasionally
B

failed to locate their own nest but eventually managed to

10.1

96.3
14.2 return to it after a few attempts. The nightly mean frequency
9.6
28
Total

3 for both Individual-A and Individual-B was 2.5 ± 0.6 and

1.7 ± 0.6 respectively.
A

Pair switching. — Pair switching behaviour is much more

(Mo), preening (Pr) defaecating (Def) and territorial display (TD). Resting (Res) is measured in minutes. Values are presented as mean ± SE.

13.5
13.5
29.8
17.5

96.9
12.7
1.3

apparent and frequently displayed by nesting pairs compared

8

to non-nesting pairs. This swapping action requires an

individual to switch its roosting position with that of its
11.3±1.4
B

1.3±0.5
1.3±0.5

0.8±0.5

partner by a transverse or alighting movement. In detail,

20±2.4
Fourth week

1±0.6

a pair is described as switching transversely by mounting

0

promptly over the partner to the other side of the nest.

Likewise, alighting is performed in the same manner through
brief fluttering but this is less commonly displayed. Nesting
A
3.3±0.9
1.8±0.6
7.3±1.9
4.8±1.3

1.8±0.5
0.8±0.3
20±2.4

swiftlets probably exhibit this behaviour to accommodate

space for each other and their nestlings (if present) inside the
0

small cup-shaped nest. During nest building, this movement

was frequently observed in the nesting pair when taking
24.1±1.8

turns secreting fresh salivary layers onto nest. From our

B
1.1±0.6
0.1±0.1
2.4±0.4
6.9±0.5

4.6±0.6
1±0.4
Third week

observations, one individual is apparently much more active

0

than the other through repetitive switching over time (A =

7.5 ± 0.9, B = 2.5 ± 0.4).
24.3±2.1
A
3.6±0.8
1.1±0.6
7.7±1.7
3.4±0.8
0.1±0.1

2.1±0.3

Parallel shifting. — Parallel shifting is a non-flight movement

3±0.8

in which the individual moves horizontally along the plank

from its roosting position to another temporary point. The
shifting movement is almost inquisitive by drifting sideways
using their clingy tarsi coupled with few wing beats on
B
0.6±0.2

6.4±1.2

5.9±0.7
0.7±0.3
27.3±2
Second week

4±1.2

the vertical beam. The distance may range from very short
to moderate (depending on the available gaps between
0

nests). With reference to the observed pair, Individual-B

was observed to move away frequently from the nest via
A
3.4±1.1
1.3±0.4

4.3±0.6
0.4±0.2

4.7±0.4
3.9±0.9
27.9±2

this movement (A = 4.3 ± 0.4, B = 6.4 ± 0.7), although at

6±1.3

very short distances. Apparently, this behaviour can be an

indicative of breaks between nest building sessions and to
give room for the partner to resume building.
24.9±2.9
3.3±1.6
4.9±1.4
1.9±0.6
3.5±1.2

1.3±0.4
B

3±0.9

Mounting. — Mounting-like contacts are less frequent

First week

compared to other roosting behaviours. Due to the relatively

0

short duration of copulation in birds (a few seconds),

the movement resembles false copulation display during
24.8±2.4
3.3±0.9
5.3±1.3
8.9±1.9

0.8±0.5

3.3±0.9
1.3±0.3

frequent body contacts inside the cramped nest. It can also

A

5±0.9

be seemingly aggressive at times with multiple attempts

at mounting and pinning down the roosting partner. More
classification

common is a single attempt at mounting immediately followed

Ethogram

by alighting beside its partner. Another scenario is the mount-

and-fly-off sequence, mostly outside the scope of view, but
RRF
PSw
PSh

Def
Mo

TD

returning to the site a few seconds to several minutes later.

PF

Pr

229
 Ramji et al.: Roosting and nesting behaviour of swiftlets

Preening. — Preening activity can be classified into two Nest building. — This particular behaviour occupied
types; self-preening and allo-preening. In general, this activity substantial hours of the overall time spent inside the swiftlet
can be considered one of the commonest behaviour performed house during the breeding period. Nightly observational
by roosting pairs. These include almost every reachable part records indicated that nest building may occur at any time
of their body (nape, back, vent, crown, underparts) which of the night but two peak activity sessions were apparent
mainly consist of their flight feathers (wing and tail feathers). before the pre-emergence period (0500–0600 hours) and
Roosting pairs typically spent several minutes (about one immediately after the returning period (1800–2000 hours).
to five minutes) in one continuous preening session. On This activity was also observed during the post-emergence
average, they spent about 25 min nightly, most frequently period (0700–1000 hours), although at a lower rate and with
during nest building intervals and immediately after returning irregular involvement between the swiftlet pair. The pair
to nest in the evening. would start the initial building by clinging firmly on the
nesting spot while the head is stretched far and held low.
Defaecating. — Observations on this particular behaviour The bill is then open and shut repeatedly while regurgitating
were not possible using the fixed IR camera, hence data were the fresh thin saliva and the process continues as the upward
collected only from the PTZ recordings on the marked pair. attachments expanded into a proper but shallow cup-shaped
Both Individual-A and Individual-B showed almost similar nest. The building process consists of chewing, retching and
patterns of defaecating. One individual might not necessarily sweeping motion from the open bill with repetitive jabs on
defaecate during the night but the mean frequencies suggest newly added layers (see Fig. 3).
that both tend to defaecate at least three to four times
while at roost. The numbers (or frequency) were relatively For close-up observation using the PTZ camera, a total of 600
constant. Mean nightly frequencies of both Individual-A hours during the nest building period was comprehensively
and Individual-B were recorded at 4.7 ± 0.4 and 5.9 ± 0.7 analysed. In this study, the marked nest was completed in 32
times respectively. days. Initial base construction begins from the first day since
nest removal until day-14. Mean total time spent during the
Territorial display. — This behavioural display can be first week for both individuals was 80.3 ± 13.7 min. From
divided into three categories; offensive, defensive and that point, the nest building rate increased substantially during
passive. An offensive territorial display is described when the second (140.2 ± 6.2 min) and third week (108.5 ± 6.1
one individual initiated a move towards another neighbouring min), subsequently slower on the fourth week (60.6 ± 6.6
individual or pair at roost, while showing some degree of min) and levelled off on the final remaining days (Table 3).
aggressiveness. In most observations, the offender was from From the cumulative data of the first week, both individuals
nearby nesting compartments. For defensive display, an showed corresponding pattern with a gradual hike before
instant reaction from this would naturally lead to retaliation reaching the peak during the second and third week (Fig.
from the harassed individual (or pair) by chasing away the 4). Both Individual-A and Individual-B started depositing
offender via shrugging and lifting both wings coupled with saliva during the first two days at relatively shorter duration
threatening jabs from the bill. This behaviour is characterised (33.6–34.4 min). The duration extended sharply by the third
by combination of rigorous pecking, wing lifting, intense day onwards, ranging from 41.5–165.7 minutes, until the
fluttering and occasionally strong quarrel-like contacts which nadir in the fourth week (day-25).
eventually forced the intruder away. Defensive actions were
commonly displayed by Individual-A in which it seemingly During the post-emergence period, Individual-A spent a
played a more protective role over its partner (A = 2.1 ± total of 104.8 min of nest building compared to just 30.5
0.3, B = 0.8 ± 0.2). Such activity showed an increasing min by Individual-B (Table 3). This accounts to13 visits
trend during the second and third week of nest building. For by the marked Individual-A compared to only six visits
the third category, passive territorial display is considered by the unmarked partner-B. Between them, Individual-A
being more sedentary towards external disturbance such as was observed building nest alone for seven sessions while
intentional or accidental approach by other swiftlet. Individual-B contributed for only a single session. Both
partners were observed building together for five sessions.
Resting. — Resting or inactive period is calculated by total Overall, Individual-A showed significant effort and endurance
of quarter-hours spent resting throughout the night and with 61.7 ± 4.3 min of mean nest-building time compared to
during the post-emergence return. During this period, swiftlet just 40.1 ± 3.4 min by Individual-B (t45 = 3.94, P = 0.001).
pairs may appear stationary (occasional head movements/ During nest building, Individual-A will normally alternate
wing lifting only) and comfortable at their roosting sites and take turn with its partner after three to five minutes
or inside their nest. Between the pair, the mean nightly of adding multiple salivary layers. We have observed that
frequency suggests that Individual-A rested longer for an the shortest duration of saliva deposition is slightly over
average of 174.8 ± 21.6 min while Individual-B for 140.8 10 seconds and the longest recorded without any break is
± 16.5 min. This was also attributed with more time spent about 7 min.
by Individual-A resting during the post-emergence period.
As the nest reached its completion, the pair was observed to
spend significantly more time resting during the final week
(A = 303.8 ± 29 min, B = 232.5 ± 41.3 min).

230
 THE RAFFLES BULLETIN OF ZOOLOGY 2013

DISCUSSION middle Baram, Sarawak, adult swiftlets have been reported

to return occasionally throughout the day to feed their
Swiftlets are free-flying birds that have a regular daily flight nestlings (Lim, 1999). However, previously unconfirmed
routine even if nesting in artificial houses. They typically observations by local birds’ nest harvesters in limestone
emerge from their nesting colony to forage for insects at caves at Bau-Jambusan in western Sarawak have claimed that
daybreak and return shortly before dark (Kang & Lee, 1991). adult black-nest swiftlets Aerodramus maximus return in the
Earlier studies have confirmed that this pattern is exhibited early morning to build the nest (Lim, pers. comm., 2000).
by most cave swiftlets (Lee & Kang, 1994; Lim, 1999). Such a claim is now proven, albeit for a different species, in
Thus a common understanding has developed that swiftlets house colonies of white-nest swiftlets using advanced video
will only return at dusk or approaching nightfall once they monitoring technology. Here, the full-daily activity patterns
have left their roosting site at dawn (Lim & Cranbrook, of the house-farmed swiftlets were expectedly similar to
2002). At Lubang Salai, a natural white-nest swiftlet cave in that of the wild colony, only with a significant time frame

Fig. 3. Ten terminologies for ethogram description of the white-nest swiftlet (A. fuciphagus). Pictorial boxes are numbered as: 1 = proximity
fluttering; 2 = random roosting flights; 3 = pair switching; 4 = parallel shifting; 5 = preening; 6 = mounting; 7 = territorial display; 8
= defaecating; 9 = resting; 10 = nest building. (i) Individual-A was observed secreting fresh salivary layers on the hinge of nest and
subsequently continued by Individual-B on day-nine. (ii) Note the head direction and bill movements in sweeping motion from side to
side. (iii) An almost complete small half-cup shaped nest observed on day-20.

231
 Ramji et al.: Roosting and nesting behaviour of swiftlets

Table 3. Nest building duration by Individual-A and Individual-B during the post-emergence period (0700–1000 hours) and the total time
spent within 25 days. Values are presented in minutes. Dash (–) indicates absent of nesting individual.

Day Post-emergence period (min) Total nest building (min) Weekly Mean (A+B) ± SE
A B A B Total (A+B)
1 3.5 0.8 21.9 11.7 33.6 First week = 80.3 ± 13.7
2 – – 21.9 12.5 34.4
3 – – 65.5 21.1 86.6
4 – – 51.9 19.8 71.8
5 – – 68.4 30.9 99.4
6 – – 82.4 46.8 129.2
7 7.7 – 63.0 43.8 106.9
8 – – 94.5 48.9 143.4 Second week = 140.2 ± 6.2
9 7.6 7.8 91.1 74.6 165.7
10 – – 76.4 46.4 122.7
11 5.5 – 73.1 53.3 126.4
12 0 – 81.5 74.7 156.2
13 4.1 – 94.4 45.8 140.2
14 – – 67.5 59.3 126.9
15 – – 72.9 43.6 116.5 Third week = 108.5 ± 6.1
16 7.1 – 68.8 52.0 120.7
17 16.8 10.9 69.7 49.0 118.7
18 17.8 – 75.8 51.7 127.5
19 16.0 4.9 58.0 37.4 95.4
20 – – 46.5 40.3 86.8
21 7.2 – 55.0 38.8 93.8
22 8.3 3.3 44.5 27.6 72.1 Fourth week = 60.6 ± 6.6
23 – 2.8 39.7 24.1 63.8
24 3.2 – 33.5 31.6 65.1
25 – – 25.6 15.9 41.5
Total 104.8 30.5 1543.4 1001.5 2544.9
Mean ± SE 4.6 ± 1.2 1.2 ± 0.6 61.7 ± 4.3 40.1 ± 3.4 101.8 ± 7.3

addition termed as the ‘post-emergence period’ (0700–1000 their roosting site due to decline of aerial prey after sunrise
hours) in between the first emergence (0600–0700 hours) compared to high abundance of insects during the early hours.
and returning period (1800–1900 hours). Nonetheless, the post-emergence return pattern observed
in this study does not seem to be linked to the availability
Our findings highlight a new discovery with reference to of food source or they taking refuge from heavy rain, but
the irregular and asynchronous visits by one partner of the rather explicitly associated to nest building by one assiduous
breeding pairs to the colony during the post-emergence partner. Unlike the black-nest swiftlet, the diet of white nest
period. Within this period, we found that significant number swiftlet is relative diverse, hence are much adaptable with
of swiftlets returned to build the nest. This behaviour has the food available in the area (Lourie & Tompkins, 2000).
been overlooked in past studies (i.e., without using CCTV
constant monitoring method) where it was assumed that Between the two swiftlet colonies, a higher returning
once the swiftlets leave their roosting place, they will spend percentage was observed in Site-I as compared to Site-II.
the entire day foraging outside except for those having a This was probably due to the experimental manipulations in
young brood to feed. Furthermore, the absence of reliable which selected colonies in Site-I was left for breeding while
24-hour power supply would not have permitted the kind of nests in Site-II were continuously harvested. As a result, the
observation undertaken in this study. Outside the caves, large high total numbers of roosting individuals counted here can
foraging flocks of swiftlets were normally most active during be attributed to the nestlings raised in each nest as opposed
early morning and late afternoon (Manchi & Sankaran, 2010). to breeding pairs without nestlings in Site-II. Nonetheless,
At this time, it is not uncommon that they will disperse and the similarity of the patterns exhibited by both colonies is the
maximize aerial hunting by following airborne insects with key point of interest. In addition, the numbers of returning
rising thermal up-currents until midday (Lim & Cranbrook, individuals during the post emergence period (i.e., to build
2002). In another study on chimney swift Chaetura pelagica, the nest) is also correlated with the swiftlet annual breeding
Zammuto & Franks (1981) suggested that swifts may re-enter cycle. Lower returning percentage (27–39%) was observed

232
 THE RAFFLES BULLETIN OF ZOOLOGY 2013

in Site-II from Apr. to Jul.2012, which coincided with the by continuous production and intermittent usage of salivary
lowest breeding activities and hence less nest building nest cement until the nest is fully completed.
activity. During this period, the swiftlets undergo an intense
annual moult cycle, whereby energy is shifted for production This was further corroborated by the close-up observation
of new feathers at the expense of reproduction (Lim, 1999). made on the marked pair. Although nest building is habitually
carried out at night, about 6.8% (Individual-A) and 3%
In the present investigation, some plausible explanations (Individual-B) from the overall nest building minutes was
that might have influenced this scenario are the high energy essentially spent during the post-emergence period (see
demands and extra investments required to build and complete Table 3). At daytime, the nest building ratio for Individual-A
the nest before the egg-laying period (Soler et al., 1998b). is three times greater than Individual-B. Hence, it is most
In addition, Lim & Cranbrook (2002) suggested that the likely that asynchrony in post emergence return between
combination of intrinsic physiological vigour and availability the nesting pair is probably associated with the breeding
of food source played an important role in determining the strategy or post-mating sexual selection process (Soler et al.,
onset of the breeding cycle in swiftlets. At the peak of the 1998a). Here, the male’s reproductive success is determined
breeding season, the salivary glands of swiftlets undergo huge through its willingness, toleration and commitment towards
production of saliva before the egg-laying period (Medway, nest building as pair-bond insurance on future parental
1962b; Lim & Cranbrook, 2002). With much enlarged and investment with the female partner (Szentirmai et al., 2005;
hyperthrophied salivary glands, an impromptu response might Alvarez & Barba, 2011). In other bird species, these male
be logical in which the swiftlets would re-enter the swiftlet quality traits in nest building behaviour were also observed
house after a quick feeding after dawn and “emptied” their in great bowerbirds Ptilonorhynchus nuchalis (Doerr, 2009),
fresh glutinous saliva load to build nest before embarking on barn swallows Hirundo rustica (Soler et al., 1998a) and
another long foraging journey. By late morning, the natural penduline tits Remiz pendulinus (Szentirmai et al., 2005).
foraging rhythm during daylight continued as they replenished
their energy in preparation for the subsequent prolonged nest On average, nest building duration for the white-nest swiftlet
building activity at night. For comparison, in broiler chicken is 30 days (Lim & Cranbrook, 2002). Our results revealed
Gallus domesticus, the salivary glands contain numerous that the nest growth rate accelerated the most during the
mucous cells which can alternate between accumulation first and second week, then gradually waned in the third
and discharge of mucous within a cycle (Newman, 2000). to fourth week and finally levelled off beyond the fourth
Studies on glycoprotein synthesis and secretion pathways week. Lim (1999) reported that nest growth curve calculated
have shown that intracellular retention half times of different from the rate of extension of the nest-cup as sigmoidal or
mammalian glycoproteins vary from 30–120 min (Yeo et reaching plateau as it entered the egg-laying phase. The
al., 1985; Bostrom et al., 1986). This is within the interval general nest building profiles for house-farmed swiftlets is
from the last nest deposition in the wee hours before dawn also similar to that of black-nest swiftlets. In accordance
until the ‘post-emergence’ return. By inference, swiftlets with Kang & Lee (1991), a single black-nest swiftlet may
seem able to maximise their salivary secretion capability spend about 25–60 min daily for nest construction and each

Fig. 4. Cumulative graph showing nest building duration (minutes) for the observed pair. Diamond-solid and square-dashed lines indicate
Individual-A and Individual-B respectively. Histogram bars denote the total minutes spent by both individuals.

233
 Ramji et al.: Roosting and nesting behaviour of swiftlets

building bout may last up to 4 min. Generally, both sexes towards other swiftlets. Another equally important activity
are known to contribute in nest construction but evidence is resting for nesting pairs. At the peak of nest building
on the comparative assiduity of labour between male and (i.e., second and third week), high energy expenditures are
female is not well understood then. Based on the nest presumably in favour of long and sufficient resting period.
building cumulative graph (Fig. 4), the marked Individual-A
was observed spending almost double nest building minutes Although the definite sexes of the marked pair were
compared to its partner. It seems that one of the sexes is unconfirmed due to wear and fading of the markings over a
much more active, laborious and diligent than the other. The 10-month period, we have strong justification that division
unequal contribution from each pair member is significant, of labour is essential between nesting pair in order to suit
which suggests that male and female swiftlets play different their annual breeding strategy. On a final note, we postulate
roles in nest building. This may be an adaptive response to that Individual-A is most likely to be the male because (i)
sustain their multi-brooded reproductive strategy in terms consistently mounting Individual-B and not the other way
of energetic. Many studies have shown that female birds round, (ii) twice more hardworking in nest building because
predominantly require enormous amount of energy for egg comparatively less energy is needed for spermatogenesis
production (e.g., Hails & Amirrudin, 1981; Hails & Turner, as compared to oogenesis (i.e., each egg is 20 × 14 mm, ≈
1985; Marvelde et al., 2012). De Neve & Soler (2002) added 2g), (iii) showing more protective behaviour over its partner
that females may also be flexible in their own reproductive as their nest reaches full size, at a time when the female is
effort by adjusting their laying date and clutch size prior known to ovulate and become receptive, and (iv) the usage
to assessing their mate’s qualities. With this observation, it of unoccupied artificial nest in swiftlet house indicates the
can be hypothesised that it is the male (i.e., Individual-A) female is perfectly comfortable to use any structure that was
that is more active in nest building because female (i.e., not constructed by her for laying.
Individual-B) swiftlets need to channel substantial energy
for egg production. Furthermore, testimonies from swiftlet
farmers confirmed that newly established breeding pairs ACKNOWLEDGEMENTS
readily used an artificial nest nailed onto the plank for laying,
thus suggesting that female swiftlet is comfortable to use any This paper is a contribution to Supplement No. 29 of the
structure deemed sufficiently large to hold her clutch without Raffles Bulletin of Zoology, marking the eightieth birthday
undergoing the laborious nest building process herself. of the Earl of Cranbrook (V). We are grateful to Dato’
Steven Lee and Datin Veronica Lau for their hospitality
Of the observed behaviour patterns, many seem to be related and granting us access into their swiftlet houses. Permission
to finding, recognising and retaining a roosting and nesting to undertake research was granted by the Sarawak Forest
position within the crowded colony (e.g., proximity fluttering, Department and Sarawak Forestry Corperation under permit
random roosting flight, pair switching, parallel shifting and no. NPW.907.4.4 (V)-105. The research was funded by the
territorial display). Others relate to self-maintenance (e.g., eScience fund-Ministry of Science, Technology & Innovation
preening, defaecating and resting) and maintenance of the (MOSTI) (Grant no. 06-01-09-SF0060) and MyBrain15-
pair bond (e.g., allo-preening and mounting). This study MyPhD scheme. We would also like to thank the staff and
also revealed that some roosting behaviours are considerably colleagues at Department of Zoology, Faculty of Resource
more evident during the nest building period than at other Science and Technology, UNIMAS.
times. Proximity fluttering, pair switching, parallel shifting
and territorial display are clearly dynamic behavioural Research on swiftlet in Sarawak was pioneered by the Earl of
modes which are useful for flight mobility, space usage and Cranbrook nearly half a century ago at Niah Cave, Sarawak in
protection in cramped and tight nesting sites. For instance, 1956. The momentum started then has never ceased, and two
higher frequencies for both proximity fluttering and random generations of Sarawakian researchers have continued Lord
roosting flight behaviours during the first week was evidently Cranbrook’s footsteps and endeavours—further dissecting
caused by frantic movements from panic reaction of the the secrets of the edible-nest swiftlets.
marked pair when their nest was experimentally removed.
As for parallel shifting, this particular behaviour probably
designate how roosting individuals locate neighbouring LITERATURE CITED
nests and perhaps aid newly breeding pairs to find suitable
or vacant spot to build their nest. This may also be a form Alvarez, E. & E. Barba, 2011. Nest characteristics and reproductive
of territorial claim as thin semi-circular salivary smears performance in great tits Parus major. Ardeola, 58: 125–136.
had been observed along both sides of an established nest. Bostrom, K., M. Wettesten, J. Boren, O. Wiklund & S. O. Olofsson,
These overlapping smears may thicken overtime but never 1986. Pulse-chase studies of the synthesis and intracellular
developed into another nest, suggesting this was not done by transport of apolipoprotein B-100 in Hep G2 cells. Journal of
a different breeding pair. Likewise for territorial display, high Biological Chemistry, 261: 13800–13806.
frequency of defensive actions was commonly displayed by Chantler, P. & G. Driessens, 1995. Swifts: A Guide to the Swifts
the marked pair in the later stages of nest building, perhaps and Treeswifts of the World. Pica Press, Sussex. 237 pp.
signalling the onset of female ovulation period. Hence in this Chazarreta, M. L., V. S. Ojeda & A. Trejo, 2011. Division of labour
case, the more prominent Individual-A is more active and in parental care in the Magellanic Woodpecker, Campephilus
protective over its partner by showing natural aggressiveness magellanicus. Journal of Ornithology, 152: 231–242.

234
 THE RAFFLES BULLETIN OF ZOOLOGY 2013

Cranbrook, Earl of & C. K. Lim, 1999. Cave swiftlets. In: Smythies, Medway, L., 1962a. The swiftlets (Collocalia) of Niah cave,
B. E. (ed. G. W. H. Davison), The Birds of Borneo, 4th Edition. Sarawak. Ibis, 104: 45–66.
Natural History Publications (Borneo), Kota Kinabalu. Pp. Medway, L., 1962b. Reproductive cycle, moult and changes in the
75–101. sublingual salivary glands of the swiftlet Collocalia maxima
De Neve, L. & J. J. Soler, 2002. Nest-building activity and laying Hume. Proceedings of the Zoological Society of London, 138:
date influence female reproductive investment in magpies: An 305–315.
experimental study. Animal Behaviour, 63: 975–980. Medway, L., 1967. The function of echo-navigation among swiftlets.
Doerr, N. R., 2009. Do male Great Bowerbirds (Ptilonorhynchus Animal Behaviour, 15: 416–420.
nuchalis) minimise the costs of acquiring bower decorations Newman, R. E., 2000. Modulation of Avian Metabolism by Dietary
by reusing decorations acquired in previous breeding seasons? Fatty Acids. Unpublished PhD thesis. University of Sydney,
Emu, 109: 237–243. Australia.
Drickamer, L. C. & S. H. Vessey, 1992. Animal Behaviour. 3rd Pechacek, P., 2005. Use of non-stop video surveillance to monitor
Edition. Wm. C. Brown Publishers. US. breeding activity of primary cavity nesters in remote areas.
Hails, C. J. & A. Amirrudin, 1981. Food samples and selectivity of Acta Ethologica, 8: 1–4.
white-bellied swiftlets Collocalia esculenta. Ibis, 123: 328–333. Pierce, A. J & K. Pobprasert, 2007. A portable system for continuous
Hails, C. J. & A. K. Turner, 1985. The role of fat and protein during monitoring of bird nests using digital video recorders. Journal
breeding in the white-bellied swiftlets (Collocalia esculents). of Field Ornithology, 78: 322–328.
Journal of Zoology, 206: 469–484. Prakash, M., A. P. Diwan, S. Srivastava & V. Singhal, 1994.
Ismail, M. Y., 1999. Social control and bird’s nest harvesting Recent Advances in Animal Behaviour. Anmol Publications
among the Idahan: A preliminary observation. Southeast Asian Pvt. Ltd., India.
Studies, 37: 3–17. Reif, V. & R. Tornberg, 2006. Using time-lapse digital video
Johnston, R. B., S. M. Bettany, R. M. Ogle, H. A. Aikman, G. A. recording for a nesting study of birds of prey. European Journal
Taylor & M. J. Imber, 2003. Breeding and fledging behaviour of of Wildlife Research, 52: 251–258.
the Chatham Taiko (Magenta petrel) Pterodroma magentae and Smithers, B. L., C. W. Boal & D. E Andersen, 2005. Northern
predator activity at burrows. Marine Ornithology, 31: 193–197. goshawk diet in Minnesota: An analysis using video recording
Kang, N. & P. G. Lee, 1991. The edible-nest swiftlets Aerodramus systems. Journal of Raptor Research, 39: 264–273.
spp. Nature Malaysiana, 16: 44–51. Soler, J. J., J. J. Cuervo, A. P. Moller & F. De Lopes, 1998a. Nest
Lee, P. G. & N. Kang, 1994. The reproductive strategis of edible-nest building is a sexually selected behaviour in the barn swallow.
swiftlet (Aerodramus spp.). The Bulletin of British Ornithologist’ Animal Behaviour, 56: 1435–1442.
Club, 114: 106–113. Soler, J. J., A. P. Méller & M. Soler, 1998b. Nest building, sexual
Lim, C. K., 1999. Sustainable Harvesting and Conservation of selection and parental investment. Evolutionary Ecology, 12:
the Edible Nest Swiftlets (Aerodramus spp.) of Sarawak. 427–441.
Unpublished PhD thesis. University of Kent at Canterbury, Szentirmai, I., J. Komdeur & T. Székely, 2005. What makes a nest-
United Kingdom. building male successful? Male behaviour and female care in
Lim, C. K., 2000. Bird’s nest soup. Wildife Conservation, February penduline tits. Behavioral Ecology, 16: 994–1000.
Issue: 30–35. Thomassen, H. A., 2005. Swift as sound. Design and Evolution of
Lim, C. K. & Earl of Cranbrook, 2002. Swiftlets of Borneo. Natural the Echolocation System in Swiftlets (Apodidae: Collocaliini).
History Publications (Borneo), Kota Kinabalu. 171 pp. Print Partners lpskamp B. V., Enschede. 220 pp.
Lourie, S. A. & D. M. Tomkins, 2000. The diets of Malaysian Tompkins, D. M.,1999. Impact of nest-harvesting on the reproductive
swiftlets. Ibis, 142: 596–602. success of black-nest swiftlets, Aerodramus maximus. Wildlife
Manchi, S. & R. Sankaran, 2010. Foraging habits and habitat use Biology, 5: 33–36.
by edible-nest and glossy swiftlets in the Andaman Islands, Yeo, K. T., J. B. Parent & K. Olden, 1985. Variability in transport
India. The Wilson Journal of Ornithology, 122: 259–272. rates of secretory glycoproteins through the endoplasmic
Mardiastuti, A., A. Mulyani & T. A. C. Gultom, 1997. Breeding reticulum and Golgi in human hepatoma cells. Journal of
success of edible nest swiftlets in man-made nesting habitat. Biological Chemistry, 260: 7896–7902.
Media Konservasi, 2: 81–83. Zammuto, R. M. & E. C. Franks, 1981. Environmental effects on
Martin, P. & P. Bateson, 1993. Measuring Behaviour. 2 Edition.
nd roosting behavior of chimney swifts. Wilson Bulletin, 93: 77–84.
Cambridge University Press, Cambridge.
Marvelde, L., S. L. Webber, H. A. J. Meijer & M. E. Visser, 2012.
Energy expenditure during egg-laying is equal for early and late
breeding free-living female great tits. Oecologia, 168: 631–638.

235
View publication stats