J Comp Physiol A (2004) 190: 233–239

DOI 10.1007/s00359-003-0489-9

O R I GI N A L P A P E R

S. Jarau Æ M. Hrncir Æ R. Zucchi Æ F. G. Barth

A stingless bee uses labial gland secretions for scent trail communication
(Trigona recursa Smith 1863)

Received: 18 September 2003 / Revised: 5 December 2003 / Accepted: 8 December 2003 / Published online: 20 January 2004

Springer-Verlag 2004

Abstract The pheromones used by several species of

stingless bees for scent trail communication are generally
Introduction
assumed to be produced by the mandibular glands. Here
The recruitment of workers to a food source is an
we present strong evidence that in Trigona recursa these
important ability of highly social insects which have
pheromones originate from the labial glands, which are
developed a multitude of ways to assemble workers in a
well developed in the heads of foragers. Analysis of the
joint effort to efficiently gather food for the entire colony
behavior involved in scent marking shows that a bee
(Wilson 1971).
extends her proboscis and rubs it over the substrate.
This also applies to the stingless bees (Hymenoptera,
A single scent marking event lasts for 0.59±0.21 s while
Apidae, Meliponini) (Lindauer and Kerr 1958, 1960;
the bee runs a stretch of 1.04±0.37 cm on a leaf.
Kerr et al. 1963; Kerr and Esch 1965; Michener 1974).
According to choice experiments the bees are attracted
Their recruitment systems range from a mere alarming
by a feeder baited with labial gland extract (84.2±6% of
of the nestmates, which stimulates them to leave the nest
the bees choose this feeder) but repelled from a feeder
and to search for food at random, to a precise com-
baited with mandibular gland extract (only 27.5±13.1%
munication of the food’s location by means of a scent
of the bees choose this feeder). They do not discriminate
trail laid by a forager between the food source and the
between two clean feeders (49.6±3% of the bees at a
nest. The species using this kind of chemical communi-
feeder). 87±5.1% of bees already feeding leave the
cation can lead large numbers of recruits to a particular
feeder after the application of mandibular gland extract
food source quickly and precisely (Lindauer and Kerr
whereas only 6.2±4.9% and 2.6±4% do so when labial
1958, 1960; Kerr and Costa Cruz 1961; Kerr et al. 1963;
gland extract or pure solvent was applied.
Johnson 1987; Schmidt et al. 2003; Jarau et al. 2003).
According to Lindauer and Kerr (1958, 1960) and
Keywords Chemical communication Æ Labial glands Æ
Kerr et al. (1963) scent marking bees run a short stretch
Scent trails Æ Stingless bees Æ Trigona
on a leaf, a blade of grass, or a small stone with their
mandibles opened. These authors therefore assumed
that the mandibular glands are the source of the trail
marking pheromones. Experimental proof for this con-
S. Jarau Æ M. Hrncir Æ F. G. Barth jecture is still lacking. Lindauer and Kerr (1958, 1960)
Biocenter, Institute of Zoology,
University of Vienna,
added mandibular gland secretions of Scaptotrigona
Althanstr. 14, 1090 Vienna, Austria postica between the marks of the scent trail previously
laid down by foragers. They found that the bees
S. Jarau (&)
Department of Experimental Ecology, ‘‘examined’’ these additional scent marks. However, re-
University of Ulm, cruits never followed a fake scent trail using mandibular
Albert-Einstein-Allee 11, gland secretions in a direction other than the bees’ nat-
89069 Ulm, Germany ural scent trail. The authors concluded that the man-
E-mail: stefan.jarau@biologie.uni-ulm.de
Tel.: +49-731-5022665
dibular gland secretions alone are not sufficient to lead
Fax: +49-731-5022683 recruits to a food source along a scent trail. Later, Kerr
et al. (1981) offered workers of T. spinipes an artificial
R. Zucchi
FFCLRP, Department of Biology,
scent trail consisting of drops of 2-heptanol, which is the
University of São Paulo, main compound of head extracts of this species,
Av. Bandeirantes 3900, 14040-901 Ribeirão Preto, SP, Brazil deposited between their nest and a feeding site where
234
about 40 bees had been trained to on the preceding day.
They concluded that the six workers that reached the
feeder during the experiment had used the 2-heptanol
trail to find it. It is very likely, however, that these bees
were not newly recruited bees but inspectors (Biesmeijer
and de Vries 2001) checking the food source they had
visited the day before.
In addition to the mandibular glands, the cephalic
portions of the labial glands are well developed in the
workers of stingless bees (Cruz Landim 1967). The
function of their secretions is not known, however (Silva
de Moraes 2002). Importantly, male bumble bees have
been shown to use the secretions of their labial glands to
scent mark flight routes and to attract virgin queens
(Kullenberg et al. 1973; Bergman and Bergström 1997).
This raises the question whether in stingless bees, close
relatives of the bumble bees (Michener 2000; Noll 2002),
the labial glands rather than the mandibular glands are
the source of the pheromones deposited by foragers
when laying scent trails.
In the present study we (1) analyze the scent marking
behavior of T. recursa and (2) examine the effect of
mandibular and labial gland extracts on bees arriving at
a set of feeders and on bees already collecting sugar
solution at the feeder.
Materials and methods
Bee nests and study site
We used two nests of T. recursa Smith 1863 collected in Ped-
regulho, SP, Brazil and brought to the Ribeirão Preto Campus of
the University of São Paulo. Experiments with nest 1 (kept in a
wooden box on the campus) were carried out between December
2000 and August 2001 and with nest 2 (kept in an observation box
inside the laboratory with an entrance tube through the wall)
in May, June, and September–December 2002.
Scent marking behavior
For each experiment (N=6) we trained one color marked bee (in
sum n=6) to a feeding site 25 m away from the nest using a feeder
containing unscented 0.5 mol l)1 sugar solution. At the experi-
mental site we replaced the training feeder by a clean one, now
containing unscented 3 mol l)1 sugar solution (experimental fee-
der). The bees never recruited newcomers nor showed any con-
spicuous behavior resembling recruitment behavior if the
concentration of the sugar solution was only 0.5 mol l)1. The first
arrival of the forager at the experimental feeder was taken as the
‘‘discovery of the food source’’ (for details on training and exper-
imental procedures see Jarau et al. 2003). We recorded the time
period and the number of visits after which the trained forager
deposited scent marks at the feeder and along the way to the nest
for the first time. We also measured the arrival time and the
number of the first newcomers at the feeder.
For the detailed analysis of how a bee deposits a scent mark we
trained a number (more than ten) of bees per filming session to a
feeding site 10 m away from the nest where the bees were allowed
to collect 3 mol l)1 sugar solution ad libitum. At a distance of
20 cm from the feeder a leaf was mounted on a stick. The bees
frequently scent marked the leaf and their behavior was filmed with
a video camera (Sony CCD-TRV 37 E) equipped with a close up
lens. The video tapes were digitized and recorded on a PC (Studi-
oDV Plus). We measured the time the bees spent with a single scent
marking event and the length of the stretch they ran on the leaf
while scent marking using the VideoPoint 2.5 soft-ware (Lenox
Softworks). The duration of a scent marking event was defined as
the time between the video frame showing the bee to first touch the
surface of the leaf and the last frame before she flew away. A total
of 56 scent marking events filmed during three sessions were ana-
lyzed in this way.
Head glands and bioassays
Mandibular and labial glands were dissected from the heads of
foraging workers under a stereo microscope and all tissues other
than the respective glands were carefully removed. For the bioas-
says we prepared extracts from 15 mandibular glands (put into
750 ll of pentane for 24 h) and from the cephalic parts of ten labial
glands (put into 500 ll of pentane for 24 h). For each bioassay we
used 10 ll of either of the two stock solutions corresponding to 0.1
bee equivalents.
Choice tests
Bees were trained to collect unscented sugar solution (50% weight
on weight) ad libitum at a training feeder 10 m away from the nest
on days when no choice experiments were performed. For the bio-
assays small plastic cups containing 50% unscented sugar solution
on plastic discs (65 mm in diameter) were used as feeders. The
distance between the test feeder (plastic disc baited with 10 ll of
mandibular or labial gland extract at the beginning of an experi-
ment) and the control feeder (10 ll pure pentane) always measured
20 cm. To avoid side bias the positions of the two feeders were
exchanged every 5 min. The set-up was mounted at the training site,
and a few drops of sugar solution were injected into the nest en-
trances at the same time in order to stimulate the foragers. The bees
quickly started to visit the known feeding site. For 20 min the
number of bees was recorded which landed and extended their
proboscises into the sugar solution on either the test feeder or the
control feeder. Bees were captured on their first visit to one of the
feeders, marked, and released at the end of the respective experiment
only. In order to avoid individual bees being tested more than once,
only those bees captured during a given experiment and not already
marked in preceding experiments were included in the analysis.
Control experiments followed the same procedure but the bees had
to choose between two identical and clean feeders.
Tests during feeding
In order to test the effect of gland extracts on bees already col-
lecting sugar solution at a feeder we trained them to visit a small
plastic dish (d=3.4 cm, h=0.8 cm) 10 m away from the nest. In
the middle and 1.5 cm above the dish a piece of filter paper
(1·1 cm) was mounted on an insect pin. As soon as several foragers
(exact number recorded in each trial) sat on the dish and collected
sugar water, 10 ll of either mandibular or labial gland extract were
slowly dropped onto the filter paper. During the subsequent 10 s
we counted the number of bees leaving the feeder. The control
experiments followed the same procedure using pure pentane. For
each test clean plastic dishes and new insect pins and filter papers
were used. Since the numbers of bees involved were not the same
for all experiments we calculated the percentage of bees that
stopped collecting and left the feeder after application of a gland
extract or pure pentane.
Statistical analyses
The data were normally distributed. Therefore parametric statisti-
cal tests could be applied. For both series of bioassays one-way
 235

ANOVA was used to check for a significant difference in the per-

centages of bees in the single groups. Multiple comparisons versus
the control group (Tukey test) were done to reveal the groups with
a significant difference in the mean percentage of bees choosing the
test or control feeder (choice tests) or in the mean percentages of
bees leaving the feeder (tests during feeding), respectively.

Results

Scent marking behavior

After replacing the training feeder by the experimental

feeder the trained foragers continued to visit the food
source regularly. Foragers always began scent marking
at the food source itself, i.e. at the feeder and the
wooden box it was placed on. The following results
are given as means±standard deviation (N=6, n=6).
The first scent marking occurred after 5.7±1.6 visits
to the experimental feeder, which was 11.2±1.6 min
after its discovery. The first newcomers arrived at the
feeder in small groups of 7.8±5.8 bees and
14.2±5.8 min after its ‘‘discovery’’ by the trained
forager, or 3.0±1.3 min after the first scent marking
had occurred.
Figure 1 shows a sequence of pictures illustrating
the scent marking by a forager of T. recursa. The bee
not only opens her mandibles but also extends her
proboscis. The proboscis is already stretched out when
the forager approaches the leaf and is rubbed on its
surface while she runs a short stretch on it. A single
scent marking event lasted on average 0.59±0.21 s
(range: 0.24–1.24 s; n=56) while the bees ran a stretch
of 1.04±0.37 cm (range: 0.41–2.00 cm; n=56) on the
leaf.

Fig. 1a–c A forager of Trigona recursa depositing a scent mark.

Mandibular and labial glands
Figure 2 illustrates the position of the cephalic portions
of the labial glands and of the mandibular glands,
respectively, within the head of a forager of T. recursa.
The labial glands were filled with secretion in 100%
(N=44) and the mandibular glands in 96.7% (N=61) of
the heads dissected when preparing the extracts used for
the choice experiments and bioassays and for further
chemical analyses of their contents (C. Schulz, unpub-
lished data).
Choice experiments and bioassays
When testing the bees’ choice between a feeder baited
with gland extract and a control feeder the foragers
behaved differently depending on whether mandibular
or labial gland extract was used. When using mandibular
gland extract the bees frequently hesitated to land, flying
forwards and backwards in front of the feeders, and
sometimes left the set-up in a rapid loop without landing
at all. When we tested the labial gland extract, on the
The bee approaches a leaf with its mandibles opened and its
proboscis extended (a), lands (b), and runs a short distance while
rubbing the proboscis on the leaf’s surface (c). Photographs taken
at three characteristic scent marking events
other hand, the bees usually approached in a slow zig-
zag flight and landed at either the test or the control
feeder without hesitation.
When the foragers had to choose between a clean
feeder and a feeder baited with mandibular gland ex-
tract, only 27.5±13.1% (N=6; n=74) landed on the
test feeder, whereas 84.2±6% (N=6; n=120) of the
bees choose the test feeder when it was treated with the
labial gland extract (Fig. 3). In the control experiments
49.6±3% (N=6; n=109) of the bees landed on one of
the two identical feeders (Fig. 3). This number is sig-
nificantly larger than the number of bees choosing the
feeder baited with mandibular gland extract and signif-
icantly smaller than the number of bees choosing
the feeder baited with labial gland extract (one-way
ANOVA, F2,17=67.813, P<0.001; Tukey comparisons,
Q(contr-mand)=6.371, P<0.05 and Q(contr-lab)=9.967,
P<0.05).
236
Fig. 2 a Head of a forager of T. recursa opened frontally and
showing the positions of the labial glands (lg) and of the
mandibular glands (mg). b A piece of a labial gland showing its
delicate structure consisting of alveoli (al) and secretion collecting
ducts (du). c The mandibular gland at the base of a mandible (md)
consists of a reservoir (re) which is partially covered by gland cells
(gc). Further abbreviations: br, brain; ce, compound eye; cl,
clypeus. Scale bar in a and c=1 mm, in b=0.1 mm
The effects of the mandibular gland extract and of the
labial gland extract on bees already collecting at the
food source differed from each other, as well (Fig. 4).
After the application of 10 ll mandibular gland extract
to the feeder 87±5.1% (N=6; n=77) of the bees flew up
within 10 s, whereas upon the application of the same
amount of labial gland extract only 6.2±4.9% (N=6;
n=77) of the bees left the feeder within 10 s. In the
control experiments only 2.6±4% (N=6; n=66) of the
bees left within 10 s following the application of 10 ll
pure pentane. There was no significant difference in the
numbers of bees leaving the feeder after the application
of labial gland extract and pure pentane, respectively,
whereas in the tests using mandibular gland extract
significantly more bees left the feeder as compared to the
control (one way ANOVA, F2,17=626.697, P<0.001;
Tukey comparisons, Q(contr-mand)=44.290, P<0.05 and
Q(contr-lab)=1.924, P>0.05).
Fig. 3 The effect of mandibular and labial gland secretions on the
foragers’ choice between two optically identical food sources. The
bars give the mean percentage of bees choosing the feeders baited
with gland extracts (gray) or one of the two identical feeders in the
control experiments (white). The whiskers give the standard
deviation, the asterisks mark a significant difference to the control
group. The numbers inside the bars give the total number of bees (n)
tested in the respective group (N=6 experiments each)
Fig. 4 The effect of mandibular and labial gland secretions on
foragers that collect sugar water at a feeder. The bars give the mean
percentage of bees that left the feeder within 10 s after 10 ll
mandibular gland extract, labial gland extract, or pure pentane
(control group) were dropped onto a filter paper placed 1.5 cm
above the center of the feeding dish. The whiskers give the standard
deviation, the asterisk marks a significant difference to the control
group. The numbers inside or above the bars give the total number
of bees (n) tested in the respective group (N=6 experiments each)
Discussion
Although nearly five decades have passed since Lindauer
and Kerr (1958, 1960) first reported that certain species
of stingless bees use scent trails as a way to communicate
the location of a food source, the mechanisms involved
in this astonishing behavior still are only poorly under-
stood. From the observation that foragers open their
mandibles when depositing a scent mark (Lindauer and
Kerr 1958, 1960) it was widely accepted that the man-
dibular glands produce the trail pheromones (e.g.,
Michener 1974; Roubik 1989), although this was never
convincingly demonstrated (see Introduction).

The laying of a scent trail by a T. recursa forager
always started at the food source proper. We therefore
tested the effect of gland extracts on the behavior of bees
newly arriving at the food source which we consider a
part of the scent trail. To successfully lead bees along an
artificial scent trail the entire way from the nest to a food
source will require knowledge about the intranidal
behavior involved in the recruitment process, i.e., in the
stimulation of the recruits by foragers to leave the nest
and to search for and follow a scent trail. These mech-
anisms are not revealed yet, although zig-zag running
and sound production by foragers inside the nest, as
reported for several scent trail laying species (Lindauer
and Kerr 1958, 1960; Esch 1967), probably play an
important role.
Labial glands used for scent marking
The findings reported in the present study strongly
suggest that in T. recursa the labial glands rather than
the mandibular glands produce the pheromones used for
scent trail communication. Our analysis of the behavior
involved during the deposition of a scent mark revealed
that a bee not only opens her mandibles while scent
marking but also extends her proboscis which is then
rubbed over the substrate. This behavior was observed
to be essentially the same when a bee left the feeder and
quickly alighted on it again to mark the food source
proper (i.e., the beginning of the scent trail) and when
she landed at twigs or leaves to mark the various spots of
the scent trail along her way back to the nest. We
therefore assume that the labial gland secretions are not
only deposited at the food source to attract other bees
but are indeed used to mark the entire scent trail. The
way by which the proboscis is rubbed over the substrate
is similar to the behavior described for bumble bee
males, which have been shown to use labial gland
secretions to scent mark plants along their flight paths
(Kullenberg et al. 1973; Bergman and Bergström 1997).
Since in T. recursa the food source proper gets always
marked by the recruiting foragers laying scent trails, we
carried out bioassays to see whether bees arriving at the
food for the first time are attracted by mandibular gland
or labial gland extracts. As shown in the control
experiments, the bees did not distinguish between two
identical feeders. By contrast, a significant majority of
the bees was attracted to a feeder baited with labial
gland extract whereas the mandibular gland extract
clearly had a repellent effect. Likewise, the application of
mandibular gland extract at the feeder scared away al-
most 87% of the bees sitting on it and collecting sugar
solution. This disturbing effect was never observed when
a forager landed to deposit a trail scent mark on the
feeder just next to other bees taking up sugar solution.
Similar to the indifference of bees already collecting to-
wards foragers depositing a trail scent mark at the food
source, they were also not disturbed by the application
of labial gland extract. The 6% of the bees that left the
237
feeder in the experiments testing the labial gland extract
did not differ significantly from the ca. 3% in the control
experiments. They probably represent those bees that
had already stopped collecting and were going to return
to the nest anyway.
The repellent effect of mandibular gland extract
Our finding of a repellent and expulsive effect of the
mandibular gland secretions agree with other studies
which showed in several species of stingless bees that
mandibular gland or head extracts (or synthetic com-
pounds previously identified from such extracts, e.g.,
citral, nerol, 2-nonanol, 2-nonanone, 2-heptanol, 2-
heptanone, octyl octanoate, octyl caproate) release
alarm behavior, aggression, or escape behavior when
presented at the nest entrance or within the nest (Blum
et al. 1970; Luby et al. 1973; Weaver et al. 1975; Johnson
and Wiemer 1982; Keeping et al. 1982; Smith and
Roubik 1983; Johnson et al. 1985; de Korte et al. 1988).
It is possible, however, that the same pheromones re-
lease different behaviors at different concentrations or in
different behavioral contexts such as nest entrance
guarding or searching for food (Wilson 1965; Blum and
Brand 1972). Blum et al. (1970) reported that workers of
T. (Geotrigona) subterranea were attracted to a small
wooden block treated with a very small drop of citral
and placed at a sucrose solution feeder. The bees grasped
the block lightly and only after releasing it they crawled
to the sugar solution and fed. Foragers also interrupted
their flight and alighted on a citral-impregnated wooden
block midway between the nest entrance and a feeder
placed only 80 cm away from it (Blum et al. 1970). The
question of whether these findings demonstrate that
citral is the primary trail pheromone employed by T.
subterranea foragers as stated by Blum et al. (1970) still
remains to be answered. Grasping the food source was
never observed in a natural foraging context. It may
have to be interpreted as aggressive behavior. The con-
centration of citral used by Blum et al. (1970) is not clear
(‘‘a very small drop’’), which makes a comparison with
our own results concerning the effect of mandibular
gland extract (0.1 bee equivalent) difficult. Certainly,
experiments using varying concentrations of gland ex-
tracts or synthetic compounds will have to be carried out
to further clarify the influence of the concentration on
the behavior released in recruited bees.
In the present study we provide evidence that the
cephalic parts of the labial glands instead of the man-
dibular glands are the source of the trail pheromones
used by foragers of T. recursa. Careful experiments are
now needed to investigate whether this is true for other
species as well or whether different species use different
glandular secretions for laying their scent trail. No study
is available yet in stingless bees on how exactly labial
gland secretions are released. According to Simpson and
Riedel (1964) honey bee workers are able to discharge
labial gland secretions even with the tongue almost
238
completely retracted. Thus, labial gland secretions
probably can be deposited on the substrate by rubbing
the dorsal surface of the almost completely retracted
tongue, which is held between the mandibles, on it.
Holding the tongue in this way could have misleadingly
been interpreted as ‘‘rubbing the mandibles on the
substrate’’ in earlier studies. In view of the new findings
the generally accepted ‘‘notion’’ that stingless bees use
their mandibular gland secretions for scent trail com-
munication should be critically reassessed.
Acknowledgements We are very grateful to Sidnei Mateus for his
valuable help in finding and collecting the bee nests in Pedregulho.
This study was supported by grant P-14328 of the Austrian Science
Foundation (FWF) to F.G.B. This work complies with the current
laws of Brazil where the experiments were carried out.
Addendum
After the submission of our manuscript a paper by Nieh
et al. (2003) appeared which has some relevance in re-
gard of our main finding. According to these authors
mandible extracts of T. hyalinata (concentration corre-
sponding to 0.5 bee equivalents) offered at a feeding
table released aggressive behavior in newly arriving bees
for about 8 min. During this time newcomers did not
land to feed. Subsequently, the bees were attracted by
the extracts. On the other hand, the authors found a
temporal synchronization of recruitment and odor
marking. Newcomers arrived at the feeder at the same
time foragers scent marked. Despite the mismatch of
these findings Nieh et al. (2003) concluded that T. hy-
alinata recruiters deposit mandibular gland secretions to
mark the scent trail.
In order to prepare the ‘‘mandibular gland’’ extract
Nieh et al. (2003) pulled out the entire mandibles of the
worker bees’ heads together with the mandibular glands
and other tissues attached to them (see their Fig. 4).
When using this method in T. recursa we frequently
pulled out parts of the labial glands together with the
mandible. A careful separation of the mandibular gland
prior to its extraction is therefore necessary.
Regarding the repellent effect of the mandibular
gland extracts the data of Nieh et al. (2003) and our own
are in agreement. The attraction of the newcomers after
eight minutes in the study by Nieh et al. (2003), however,
probably was due to labial gland secretion unknowingly
contained in the ‘‘mandibular gland’’ extract.
References
Bergman P, Bergström G (1997) Scent marking, scent origin, and
species specificity in male premating behavior of two Scandi-
navian bumblebees. J Chem Ecol 23:1235–1251
Biesmeijer JC, Vries H de (2001) Exploration and exploitation of
food sources by social insect colonies: a revision of the scout-
recruit concept. Behav Ecol Sociobiol 49:89–99
Blum MS, Brand JM (1972) Social insect pheromones: their
chemistry and function. Am Zool 12:553–576
Blum MS, Crewe RM, Kerr WE, Keith LH, Garrison AW, Walker
MM (1970) Citral in stingless bees: isolation and functions in
trail-laying and robbing. J Insect Physiol 16:1637–1648
Cruz Landim C da (1967) Estudo comparativo de algumas glând-
ulas das abelhas (Hymenoptera, Apoidea) e respectivas impli-
cações evolutivas. Arq Zool S Paulo15:177–290
Esch H (1967) Die Bedeutung der Lauterzeugung für die Vers-
tändigung der stachellosen Bienen. Z Vergl Physiol 56:199–220
Jarau S, Hrncir M, Schmidt VM, Zucchi R, Barth FG (2003)
Effectiveness of recruitment behavior in stingless bees (Apidae,
Meliponini). Insectes Soc 50:365–374
Johnson LK (1987) Communication of food source location by the
stingless bee Trigona fulviventris. In: Eder J, Rembold H (eds.)
Chemistry and biology of social insects. Peperny, München, pp
698–699
Johnson LK, Wiemer DF (1982) Nerol: an alarm substance of the
stingless bee, Trigona fulviventris (Hymenoptera: Apidae).
J Chem Ecology 8:1167–1181
Johnson LK, Haynes LW, Carlson MA, Fortnum HA, Gorgas DL
(1985) Alarm substances of the stingless bee, Trigona silvestri-
ana. J Chem Ecology 11:409–416
Keeping MG, Crewe RM, Field BI (1982) Mandibular gland
secretions of the old world stingless bee, Trigona gribodoi
Magretti: isolation, identification, and compositional changes
with age. J Apic Res 21:65–73
Kerr WE, Costa Cruz C da (1961) Funções diferentes tomadas pela
glândula mandibular na evolução das abelhas em geral e em
‘‘Trigona (Oxytrigona) tataira’’ em especial. Rev Brasil Biol
21:1–16
Kerr WE, Esch H (1965) Comunicação entre as abelhas sociais
brasileiras e sua contribuição para o entendimento da sua
evolução. Ciên Cult (Sao Paolo) 17:529–538
Kerr WE, Ferreira A, Mattos NS de (1963) Communication among
stingless bees—additional data (Hymenoptera: Apidae). J N Y
Entomol Soc 71:80–90
Kerr WE, Blum M, Fales HM (1981) Communication of food
source between workers of Trigona (Trigona) spinipes. Rev
Brasil Biol 41:619–613
Korte M de, Weissenbacher KH, Crewe RM (1988) Chemical
signals in a stingless bee Trigona (Meliplebeia) denoiti Vachal
(Hymenoptera: Apidae: Meliponinae). J Entomol Soc South
Afr 51:9–16
Kullenberg B, Bergström G, Bringer B, Carlberg B, Cederberg B
(1973) Observations on scent marking by Bombus Latr. and
Psithyrus Lep. males (Hym., Apidae) and localization of site of
production of the secretion. Zoon [Suppl] 1:23–30
Lindauer M, Kerr WE (1958) Die gegenseitige Verständigung bei
den stachellosen Bienen. Z Vergl Physiol 41:405–434
Lindauer M, Kerr WE (1960) Communication between the workers
of stingless bees. Bee World 41:29–41, 65–71
Luby JM, Regnier FE, Clarke ET, Weaver EC, Weaver N (1973)
Volatile cephalic substances of the stingless bees, Trigona
mexicana and Trigona pectoralis. J Insect Physiol 19:1111–1127
Michener CD (1974) The Social behavior of the bees: a comparative
study. Harvard University Press, Cambridge, Massachusetts
Michener CD (2000) The bees of the world. Johns Hopkins Uni-
versity Press, Baltimore
Nieh JC, Contrera FAL, Nogueira-Neto P (2003) Pulsed mass
recruitment by a stingless bee, Trigona hyalinata. Proc R Soc
London Ser B 270:2191–2196
Noll FB (2002) Behavioral phylogeny of corbiculate Apidae
(Hymenoptera; Apinae), with special reference to social
behavior. Cladistics 18:137–153
Roubik DW (1989) Ecology and natural history of tropical bees.
Cambridge University Press, Cambridge
Schmidt VM, Zucchi R, Barth FG (2003) A stingless bee marks the
feeding site in addition to the scent path (Scaptotrigona aff.
depilis Moure 1942). Apidologie 34:237–248
Silva de Moraes RLM (2002) Glândulas salivares do adulto.
In: Cruz Landim C da, Camargo Abdalla F (eds) Glândulas

Exócrinas das Abelhas. FUNPEC—Editora, Ribeirão Preto
São Paulo, pp 51–70
Simpson J, Riedel IBM (1964) Discharge and manipulation of la-
bial gland secretion by workers of Apis mellifera (L.) (Hyme-
noptera: Apidae). Proc R Entomol Soc London Ser A 39:76–82
Smith BH, Roubik DW (1983) Mandibular glands of stingless bees
(Hymenoptera: Apidae): chemical analysis of their contents and
biological function in two species of Melipona. J Chem Ecol
9:1465–1472
239
Weaver N, Weaver EC, Clarke ET (1975) Reactions of five species
of stingless bees to some volatile chemicals and to other species
of bees. J Insect Physiol 21:479–494
Wilson EO (1965) Chemical communication in the social insects.
Science 149:1064–1071
Wilson EO (1971) The insect societies. Belknap Press of Harvard
University Press, Cambridge, Massachusetts