Content uploaded by Márcia Motta Maués
Author content
All content in this area was uploaded by Márcia Motta Maués
Content may be subject to copyright.
Hindawi Publishing Corporation
Psyche
Volume 2012, Article ID 978019, 9pages
doi:10.1155/2012/978019
Research Article
Pollination Requirements and the Foraging Behavior of
Potential Pollinators of Cultivated Brazil Nut
(
Bertholletia excelsa
Bonpl.) Trees in Central Amazon Rainforest
M. C. Cavalcante,1F. F. Oli ve i r a ,2M. M. Mau´
es,3and B. M. Freitas1
1Department of Animal Science, Federal University of Cear´
a (UFC), Avenida Mister Hull 2977, Campus do Pici, CEP 60021-970,
Fortaleza, CE, Brazil
2Department of Zoology, Federal University of Bahia (UFBA), Rua Bar˜
ao de Geremoabo 147, Campus de Ondina, CEP 40170-290,
Salvador, BA, Brazil
3Entomology Laboratory, Embrapa Amazˆ
onia Oriental (CPATU), Travavessa Dr. En´
eas Pinheiro s/n, CEP 66095-100,
Bel´
em, PA, Brazil
Correspondence should be addressed to B. M. Freitas, freitas@ufc.br
Received 5 December 2011; Revised 7 March 2012; Accepted 25 March 2012
Academic Editor: Tugrul Giray
Copyright © 2012 M. C. Cavalcante et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
This study was carried out with cultivated Brazil nut trees (Bertholletia excelsa Bonpl., Lecythidaceae) in the Central Amazon
rainforest, Brazil, aiming to learn about its pollination requirements, to know the floral visitors of Brazil nut flowers, to investigate
their foraging behavior and to determine the main floral visitors of this plant species in commercial plantations. Results showed
that B. excelsa is predominantly allogamous, but capable of setting fruits by geitonogamy. Nineteen bee species, belonging to
two families, visited and collected nectar and/or pollen throughout the day, although the number of bees decreases steeply after
1000 HR. Only 16, out of the 19 bee species observed, succeeded entering the flower and potentially acted as pollinators. However,
due to the abundance, flower frequency and foraging behavior of floral visitors, it was concluded that only the species Eulaema
mocsaryi and Xylocopa frontalis could be considered relevant potential pollinators.
1. Introduction
Brazil nut (Bertholletia excelsa Bonpl., Lecythidaceae) is na-
tive from the Amazon forest occurring in the wild from 5◦N
to 14◦S in Venezuela, Colombia, Peru, Bolivia, Suriname,
Guyana, and Brazil [1–3]. It is harvested for its nut, which is
extracted from inside the large, rounded and hard-to-break
fruit collected on the ground after falling from the trees
[4]. Most production is for export comprising an important
source of food and income to the indigenous people [5].
Brazil nut is believed to be an allogamous species present-
ing mellitophilous pollination syndrome, thus depending on
biotic pollinators to set fruits [6]. However, little is known
about its breeding system and pollination requirements. The
blooming period occurs from September to December, peak-
ing in November, and flowers are produced profusely in
vertical terminal panicles [6,7]. The flower is large (c.a.
3.9 cm in length ×3.6 cm in width), zygomorphic, with two
to three sepals, and six yellowish petals [6,8]. It bears a
curled hood made of congruent staminodes, called ligule,
that in association with the petals form a chamber which
conceals stamens, stigma, and nectaries [8,9]. The large size
and strength of the hood restricts and selects flower visitors
to medium- and large-sized bees strong enough to uncurl
it [7,8]. Anthers begin to dehisce while the flower is still
closed, around 0100 HR-0130HR and over 90% of anthers
are shedding pollen by 0300 HR. Pollen viability ranges from
76% to 86.5% and remained viable until 1400 HR [10,11].
Anthesis takes place between 0430 HR to 0500 HR, and petals
fall offafter 24 h. When fecundation does not occur, the
pistil drops after 48 h [10]. The ovary bears an average of 20
ovules, and only 0.28 to 0.40% of the flowers produced set
2Psyche
fruits [12,13]. Fruits take an average of 15 months to mature
[7,14].
There are few studies investigating floral visitors of Brazil
nut, and usually they are restricted to the genus level.
Prance and Mori [15] stated that the main pollinators of
species belonging to the Lecythidaceae family are Bombus
and Euglossa bees.M¨
uller et al. [10], dealing with B. excelsa,
believe that large-sized bees of the genus Bombus are the
main pollinators of this species, while a study carried out in
Bolivia, suggested that euglossine bees are the effective
pollinators [13]. However, a study carried out in the state of
Acre, Brazil, points out to bees of the genus Xylocopa [16].
Only Nelson et al. [9] in a study nearby the city of Manaus,
State of Amazonas, and Mau´
es [7], working close to the city
of Bel´
em, State of Par´
a, have identified the bee species visiting
Brazil nut flowers to the species level. In both cases, they were
all medium-to large-sized bees: Eulaema seabrai (Moure,
1960), Epicharis rustica (Olivier, 1789), Ep. umbraculata
(Fabricius, 1804), Eulaema nigrita (Lepeletier, 1841), El.
cingulata (Fabricius, 1804), in Nelson et al. [9]work,and
Xylocopa frontalis (Olivier, 1789), X. aurulenta (Fabricius,
1804), Ep. rustica (Olivier, 1789), Ep. affinis (Smith, 1874),
Centris similis (Fabricius, 1804), El.nigrita, El. cingulata,
Bombus brevivillus (Franklin, 1913), and B. transversalis
(Olivier, 1789), in Mau´
es [7] report. Recently, Santos and
Absy [17]reportedX. frontalis and El. mocsaryi (Friese,
1899), as the most abundant floral visitors of B. excelsa
flowers in Itacoatiara county, State of Amazonas.
There is a lack of precise information on the breeding
system and floral visitors of B. excelsa.Thisworkaimedto
investigate the pollination requirements, learn about the
identity and foraging behavior of visitors to Brazil nut
flowers, and discuss their potential as pollinator of this plant
species. Such knowledge is remarkably important in develop-
ing policies of sustainable use of the forest and conservation
of the native bee pollinators. It may also help to explain and
to overcome the low productivity observed in commercial
plantations of Brazil nut [8–10].
2. Methods
The experiment was carried out in Aruan˜
a farm, situated on
the road Manaus-Itacoatiara, km 215, county of Itacoatiara,
State of Amazonas, Brazil, at 3◦030.63 Sand58
◦50
1.50 W. The farm total area comprises 12,000 ha, of which
3,600 are cultivated with 20 varieties of grafted Brazil nut
trees. The trees are spaced at 20 ×20 m reaching approxi-
mately 1,300,000 trees. It is the largest Brazil nut plantation
in the world.
Four trees (three belonging to variety 609 and one to
variety Abufari) were chosen at random out of those in
blooming. These trees were ca. 0700 HR apart from each
other and ranged from 25–30 m in height. Scaffolds were
built by the side of each tree, allowing to spot visually 60%
of their canopies and access flowers for data collection. Field
observations were carried out for 78 days, from October
to December 2007, covering the whole flowering period,
especially its peak in November.
2.1. Pollination Requirements. Aiming to know the pollina-
tion requirements of Brazil nut trees and the role of bees in
pollinating this plant species, we applied five pollination
treatments to the trees during their blooming.
T1: Open Pollination. We marked 655 buds with satin
threads tied to their petiole in the day before flower anthesis.
These buds were observed throughout the anthesis and
flower lifespan until they have fallen from the trees or being
set, until 25 days later. In this treatment, we aimed to know
the natural levels of pollination of Brazil nut trees in the area
studied.
T2: Restricted Pollination. 326 buds were covered with mus-
lin bags and remained bagged for 25 days. The aim of this
treatment was to verify the dependence or nondependence
of Brazil nut flowers on biotic pollination.
T3: Hand Cross-Pollination. 150 buds were marked with
satin threads and bagged with muslin bags. Next day, after
anthesis, flowers were unbagged and manually pollinated
with pollen grains from flowers of another Brazil nut
tree being deposited directly on the stigma. Donor flowers
were collected minutes before we start to perform hand
pollination and taken immediately to receptor tree. Then,
pollen grains were removed from the anthers of the donor
flower using a fine painting brush and transferred promptly
to the stigma of the receptor flower. Immediately after
hand-pollinated, the flowers were protected with muslin
bags for 25 days. This treatment indicates cross-pollination
requirements of the brazil nut tree and the existence any
pollination deficit by comparison to natural fruit set in the
area (open pollination).
T4: Hand Self-Pollination. We marked 98 buds and followed
the same procedure described above, except that pollen
grains were transferred between anthers and stigma of the
same flowers. In this treatment, results show if the Brazil nut
tree is self-compatible or not.
T5: Geitonogamy. The same procedure above was repeated
here with 78 buds, but pollen grains were transferred from
anthers of a flower to the stigma of a different flower from the
same tree. We aimed to learn if the Brazil nut tree shows
any sort of incompatibility, this kind of crossing and, its
dependence on foreign pollen grains.
In this experiment, colors of the satin threads varied
according to the treatment, and satin threads were carefully
tied to the buds’ petiole avoiding damaging the buds,
obstruction of the anthesis, and normal development of
the flower and fruit set. Also, all hand pollinations were
performed between 0600 HR and 0800 HR when, according
to M¨
uller et al. [10], fecundity is greatest.
Brazil nut fruits take an average of 14 months to ripe,
and other factors besides pollination can interfere with fruit
persistence on trees [7,14]. Thus, in all tests we assessed
initial fruit set 25 days after flower manipulation as a measure
of pollination effectiveness. This is a reliable measure because
unpollinated flowers fall from the trees in the same day they
Psyche 3
open, while pollinated ones remain on the trees and show an
ovaryabout1.5mmindiameter25dayslater.
2.2. Floral Visitors and Foraging Behavior. Samples of all
floral visitors were collected from each tree using entomo-
logical nets at every hour from 0500 HR to 1700 HR. Then,
insects were killed in a lethal chamber with ethyl acetate,
pinned, identified at species level and, sexed, and counted to
determine their specific abundance.
During blooming, the foraging behavior of each flower
visiting species was recorded considering the following pa-
rameters: frequency, abundance, hour of the day and number
of visits, time spent per flower, approach and handling of
the flower, and entry to the flower. Data were collected
using a notepad, a stop watch, a video and photo camera
Sony Cyber-shot DSC-H50 9.1MP, and by means of visual
observation of the bees foraging on the flowers, most of
them are out of the reach of the observer but in his sight.
Recording was initiated when the bee species arrived to
the tree and stopped when the insect flew away or went
out of the observer’s sight, that was limited to only part
of the canopy. All data were collected in 25 periods of 30
minutes each, starting at 0500 HR and ending by 1700 HR.
This information was later related to temperature, and air
relative humidity records obtained every 30 minutes using
a digital thermal hygrometer, model Impac TH02, because
there are evidences that increases in ambient temperature
have a negative impact o the foraging of bees [18,19].
2.3. Statistical Approach. Data on pollination requirements
did not conform to the ANOVA presumptions due to their
binomial character (set fruit or nonset fruit) and were an-
alysed using the nonparametric test of Kruskall-Wallis, and
means were compared by the nonparametric Dunn’s test.
Data regarding the number of flowers visited per tree and
time spent per flower were analysed by ANOVA, and means
were compared a posteriori by Tukey test at 5%. All tests were
performed using SPSS 19 Statistics program.
3. Results
3.1. Pollination Requirements. There were significant (P<
0.05) differences between treatments for fruit set (Ta b l e 1).
The hand cross-pollination treatment set the greater number
of fruits and differed (P<0.001; KW =54.295) from all
other treatments, while the geitonogamy treatment did not
differ (P<0.001, KW =54.295) to the free pollination
treatment. Flowers submitted to the restricted and hand self-
pollination treatments set no fruits (Ta b l e 1).
3.2. Flower Visitors and Foraging Behavior. Flowers of B.
excelsa were visited by a wide range of animals, such as
Hymenoptera (bees), Lepidoptera (butterflies and moths),
and birds (hummingbirds). In Hymenoptera, a great variety
of bee species was observed and collected visiting Brazil nut
flowers. These bees belonged to two families (Apidae and
Megachilidae) in a total of 19 species (Tabl e 2 ).
Observations on the foraging behavior of floral visitors
and potential pollinators showed that bees collect both
Tab l e 1: Initial fruit set of Brazil nut (Bertholletia excelsa)flowers
submitted to five pollination treatments: open pollination, bagged
with muslin bags, hand cross-pollination, hand self-pollination,
and geitonogamy. Itacoatiara, Amazonas, Brazil, 2007.
Treatments nFruit set
(number) Fruit set (%)
Free pollination 655 20 3.05b
Pollinator exclusion 326 0 0
Hand cross-pollination 159 29 19.33a
Hand self-pollination 98 0 0
Geitonogamy 78 3 3.85b
∗Values followed by the same letters are not significantly different
(P<0.001; Kruskal-Wallis ANOVA).
pollen and nectar from B. excelsa flowers. The place from
where bees collected nectar from the flowers varied according
to the species size. Larger bee species harvested nectar from
the ligule base, while smaller species got inside the flower to
collect the nectar present at the base of the anthers.
Bees initiated harvesting pollen and nectar at 0515 HR
and reached a peak of foraging activity between 0530 HR
and 0600 HR. After 1000 HR the number of bees foraging
on flowers dropped steeply, coinciding to the temperature
increase and relative air humidity drop (Figure 1). However,
a small number of bees kept foraging in the afternoon, spe-
cially the species Xylocopa frontalis. On the contrary of
M¨
uller et al. [10] report of bees starting to forage earlier in
the dawns following full moon nights, we did not register any
difference from the other nights (n=2).
The most abundant floral visitor of Brazil nut was the
carpenter bee Xylocopa frontalis. This species was the first one
to arrive at the flowers (around 0515 HR) to collect nectar
and pollen (Figure 2(a)) and was found in great numbers
and frequency throughout the whole blooming season of the
trees studied. After reaching a flower, X. frontalis was used
to make a brief inspection of it and, if not rejected, pushed
inside the flower using its ligule as a platform to collect
nectar from the base of the ligule itself. This bee species was,
apparently, the one which carried more pollen on its body,
especially on the back of the thorax, head, and in the scopa.
A typical behavior observed in X. frontalis while foraging
was to sit on a flower and groom pollen out of its body
towards the scopa and discard with the forelegs the exceeding
pollen grains. Xylocopa frontalis was among the three bee
species that visited most flowers per tree and spent over than
10 seconds per visit (Tab l e 3 ). Males were observed visiting
flowers for nectar, but they also carried great amounts of
pollen on their thorax (Figure 2(b)).
Centris denudans (Lepeletier, 1841) was observed visiting
flowers (Figure 2(h)) in the canopy of all trees of this study.
It was present throughout the blooming season, carrying
small amounts of pollen on the back of the thorax, despite
the bee large size. This species frequently chased after other
individuals of the same species in quick flights over the
canopy, possibly to drive the other bee offthe food source
or to mate with her. It was one of the few species observed
foraging in the afternoon, the hottest part of the day,
4Psyche
Tab l e 2: List of families, species, sex and body size of bees, floral visitors, and potential pollinators of Brazil nut (Bertholletia excelsa),
collected in a commercial cultivation in the county of Itacoatiara, state of Amazonas, Brazil, 2007.
Family Species Sex Body size (mm)
±s.d.
Apidae Xylocopa (Neoxylocopa)frontalis (Olivier, 1789) ♂♀ 34.60 ±0.10
Apidae Epicharis (Epicharana)flava (Friese, 1900) ♀17.40 ±0.26
Apidae Epicharis (Epicharana)conica (Smith, 1874) ♂♀ 12.30 ±0.97
Apidae Epicharis (Epicharis)umbraculata (Fabricius, 1804) ♀28.70 ±1.10
Apidae Epicharis (Parepicharis)zonata (Smith, 1854) ♀15.20 ±0.75
Apidae Centris (Ptilotopus)americana (Klug, 1810) ♀35.10 ±0.88
Apidae Centris (Trachina)carrikeri (Cockerell, 1919) ♂5.50 ±1.04
Apidae Centris (Xanthemisia)ferruginea (Lepeletier, 1841) ♀7.80 ±0.45
Apidae Centris (Ptilotopus)denudans (Lepeletier, 1841) ♂♀ 34.20 ±1.75
Apidae Eulaema (Eulaema)meriana (Olivier, 1789) ♂♀ 33.40 ±1.20
Apidae Eulaema (Apeulaema)mocsaryi (Friese, 1899) ♂♀ 15.60 ±0.84
Apidae Eulaema (Apeulaema)cingulata (Fabricius, 1804) ♀14.60 ±0.93
Apidae Bombus (Fervidobombus)transversalis (Olivier, 1789) ♀16.40 ±2.86
Apidae Eufrisea purpurata (Mocs´
ary, 1896) ♀10.80 ±0.89
Apidae Eufrisea flaviventris (Friese, 1899) ♀15.30 ±1.33
Apidae Apis mellifera scutellata (Lepeletier, 1836) ♀4.40 ±0.19
Apidae Frieseomelitta longipes (Smith, 1854) ♀1.50 ±0.24
Apidae Melipona (Michmelia)lateralis (Erichson, 1848) ♀4.90 ±0.32
Megachilidae Megachile sp. 1 ♀4.65 ±0.76
0
10
20
30
40
50
60
70
80
90
100
Values
Time of the day
Bees
Humidity (%)
Temperature (◦C)
5:00–5:30
6:00–6:30
7:00–7:30
8:00–8:30
9:00–9:30
10:00–10:30
11:00–11:30
12:00–12:30
13:00–13:30
14:00–14:30
15:00–15:30
16:00–16:30
17:00–17:30
Figure 1: Frequency of floral visitors associated to temperature and
relative humidity (at each 30 minutes) in a commercial cultivation
of Brazil nut (Bertholletia excelsa) in the county of Itacoatiara, state
of Amazonas, Brazil, 2007.
although most of its foraging activities were recorded in
the morning. This bee species approached the flowers in
adifferent way of X. frontalis because it did not inspect
and rarely rejected a flower, entering the flower immediately
after reaching it, but also harvested nectar from the ligule
base. Centris denudans ranked second among the species
that visited most flowers per tree, usually flowers close to
each other, and also spent over than 10 seconds per flower
visit (Tab l e 3). Males were observed and recorded visiting B.
excelsa flowers, and mating events on Brazil nut flowers were
also registered.
Eulaema meriana (Olivier, 1789) was also present
throughout the blooming season, but only in the morning.
Like X. frontalis, frequently rejected some flowers but always
carried large amounts of pollen in its corbicula. Due to
its large glossa, this bee species also collected nectar from
the ligule base landing on the ligule itself (Figure 2(k)).
El. meriana was the bee species that visited most flowers
per tree, usually neighboring flowers, spending over than
16 seconds per visit (Tab l e 3 ). Males of this species were
observed harvesting nectar from the Brazil nut flowers.
Centris americana (Klug, 1810) was seen only in some
moments of the blooming period and always in small num-
bers and low frequency to flowers, never exceeding one
individual per tree at a given time. This species approached
the flower like the other large-sized bees, using the ligule
as a platform for landing and collecting nectar from the
ligule base (Figure 2(j)). It spent less than 8 seconds per visit
(Tab l e 3 ).
Bombus transversalis was recorded only in the beginning
of the blooming season (Figure 2(o)). It was one of the
species that spent most time per flower visit, reaching up
to 90 seconds inside a flower in some visits. Despite staying
long in the flower, B. transversalis usually transported small
Psyche 5
(a) (b) (c) (d)
(e) (f) (g) (h)
(i) (j) (k) (l)
(m) (n) (o) (p)
(q) (r)
Figure 2: Approach to flowers of Brazil nut (Bertholletia excelsa) by distinct bee species in a commercial cultivation in the county of
Itacoatiara, state of Amazonas, Brazil, 2007. ((a); (b)) Xylocopa frontalis (♀and ♂, resp.); (c) Epicharis (Epicharana) flava (♀); ((d); (e))
Epicharis (Epicharana) conica (♀and ♂, resp.); (f) Epicharis (Epicharis) umbraculata (♀); (g) Epicharis (Parepicharis) zonata (♀); (h) Centris
(Ptilotopus) denudans (♀); (i) Centris ferruginea (♀); (j) Centris (Ptilotopus) americana (♀); (k) Eulaema (Eulaema) meriana (♀); ((l); (m))
Eulaema (Apeulaema)mocsaryi (♂and ♀, resp.); (n) Eulaema (Apeulaema)cingulata (♀); (o) Bombus (Fervidobombus)transversalis (♀); (p)
Eufriesea flaviventris (♀); (q) Megachile sp.1; (r) Frieseomelitta longipes robbing pollen from El. (A.) mocsaryi.
amounts of pollen and visited only a few flowers per tree
(Tab l e 3 ). Due to its medium size, this species entered almost
entirely in the flower to collect nectar at the ligule base.
Eulaema mocsaryi was the second most abundant and
frequent species over the whole blooming season, mainly in
the morning shift (Figure 2(m)) but also observed visiting
flowers in the afternoon. It frequently rejected flowers that
possibly had been previously visited by other bee. Between
two flower visits, while in flight or landing on a leaf,
individuals of this species combed pollen from their bodies
into the corbicula making large pollen loads. This bee visited
less than five flowers per tree moving quickly to other trees
6Psyche
Tab l e 3: Bee relative abundance, mean number (±standard error: SE) of flowers visited per tree by ten bee species and mean time (±
standard error: SE), in seconds, spent by twelve bee species per visit to flowers of Brazil nut (Bertholletia excelsa) variety 609, under cultivation
in the Amazon rainforest (n: number of bees recorded per species).
Species
Relative
abundance Number of flower visits per tree Time spent per flower visit
(%) nX±S.E. nX±S.E.
Xylocopa frontalis 62.85 136 11.33 ±0.834 abc 64 11.63 ±0.754 bcd
Centris denudans 6.84 35 14.71 ±2.368 bc 64 11.96 ±0.736 bcd
Centris americana 1.11 — 4 7.73 ±0.694 cd
Centris ferruginea 0.55 3 3.67 ±2.667 c 31 9.14 ±0.854 cd
Eulaema meriana 6.65 17 15.10 ±2.358 a 57 16.05 ±1.204 bc
Eulaema mocsaryi 12.20 72 4.36 ±0.514 abc 48 15.34 ±1.488 bc
Eulaema mocsaryi (male) 98.33 ±2.677 abc 55 5.68 ±0.265 d
Epicharis conica 3.88 8 2.75 ±0.773 c 7 18.39 ±2.714 bcd
Epicharis flava 0.37 7 4.43 ±1.288 bc 45 11.86 ±1.354 bcd
Epicharis zonata 0.92 9 1.67 ±0.289 c 3 31.38 ±13.090 a
Eufrisea flaviventris 0.37 6 7.33 ±3.373 abc 58 5.96 ±0.983 d
Eufrisea purpurata 0.74 — 4 14.54 ±5.809 bcd
Bombus tranversalis 3.51 3 6.33 ±2.963 abc 42 27.61 ±1.928 a
∗Values followed by the same letters are not significantly different (P<0.005; ANOVA).
(Tab l e 3 ). However, when visiting a flower, El. mocsaryi spent
over 15 seconds increasing the chance to deposit pollen on
the stigma (Tab l e 3).
Epicharis conica (Smith, 1874) was present throughout
the blooming season and like El. mocsaryi was more frequent
in the morning shift, but also present in the afternoon. Due
to its small size, this species penetrates the flower almost
entirely and unlike the previous species present here, the bee
makes a turn inside the flower before leaving it facing out
(Figure 2(d)). This bee was the second species that visited
less flowers per plant, but took over 18 seconds per visit
(n=7) (Tab l e 3 ). Males also visited flowers and pushed
their bodies completely through the petals getting hidden
by the ligule while inside the flower (Figure 2(e)). Because
of this behavior, their presence was only noticed because the
buzzing noisy produced when approaching the flower.
Epicharis flava (Friese, 1900) was present in reduced
numbers and only when most trees were in bloom. It carried
much pollen on the back of the thorax (Figure 2(c)),
outstanding as a potential pollinator of Brazil nut flowers.
This bee visited few flowers per tree and spent around 12
seconds per visit (Tab l e 3 ).
Epicharis zonata (Smith, 1854) is a small bee that like
other species of its size gets inside the flower becoming
hidden from sight and leaves it facing out carrying small
amounts of pollen on its body (Figure 2(g)). This bee was
only found in the peak of the blooming season, mainly
around 0900 HR. It is a fast-flying bee that moves between
trees frequently making difficult to track its path over a
single tree canopy. As a consequence, Ep. zonata produced
the smaller number of flowers visited per tree among all bee
species observed in this study, compensated for the longest
period of time registered for flower visit (Tab l e 3 ).
Eufriesea flaviventris (Friese, 1899) is a medium-sized,
fast-flying species, and the faster flower visitor observed in
this study spending around only six seconds per visit (Table
3), but many times revisiting consecutively the same Brazil
nut flower. This was the only species observed to collect
exclusively pollen (Figure 2(p)). It also rejected flowers
previously visited by other bees and combed the pollen from
its thorax to the corbicula while in flight.
Centris ferruginea (Lepeletier, 1841) is a fast-flying,
small-sized bee that penetrates the flower almost entirely
using the ligule as a platform. It also leaves the flower facing
out (Figure 2(i)) and carrying small amounts of pollen on
the back of the thorax. Usually was only noticed due to the
buzzing sound of its flight over the canopy. This bee species
also visited few flowers per tree, favoring cross-pollination
(Tab l e 3 ).
Megachile sp. was the smaller species registered visiting
Brazil nut flowers in this study. It penetrated entirely the
flower pushing its body among the petals and ligule and also
left the flower facing out with small amounts of pollen on
its ventral scopa (Figure 2(q)). Because of its size, probably
collected nectar from the base of the anthers and stigma,
although it s not possible to know for sure because the bee
remained hidden inside the flower while sipping nectar. Due
to its low frequency and high flight speed, only one visit was
registered.
Eulaema cingulata,Epicharis umbraculata,Centris car-
rikeri (Cockerell, 1919), and Eufriesea purpurata (Mocs´
ary,
1896) were collected and observed visiting Brazil nut flowers;
however, only in rare occasions not allowing even photos to
be taken for the two latter species.
Melipona lateralis (Erichson, 1848) was seen only once
visiting a flower and captured immediately after leaving the
Psyche 7
flower. No further sights were possible until the end of the
study.
Apis mellifera scutellata (Lepeletier, 1836) was the only
nonnative species recorded in this study, constituting an
invading bee in the Amazon ecosystem. It was present in
small numbers flying over the canopy, mainly early in the
morning. Because of its small size and strength, the bee could
not pull the ligule back as a platform as did the larger bee
species or push herself among the ligule and petals to get
inside the flower as done by other medium and small-sized
bees and remained flying over the flowers and landing to
collect small amounts of pollen fallen on petals or ligule after
the visits by larger bees.
Frieseomelitta longipes (Smith, 1854) was found in the
trees all over the morning shift and in greater numbers
than A. mellifera and, for the same reasons, also did not get
assessment of the floral resources inside the flower. However,
F. longipes showed the behavior of trying to rob pollen from
the corbicula of large bees in the moment they were visiting
the flowers (Figure 2(r)), sometimes making these bees to
give up the flower.
Besides bees, butterflies, hawk moths, and humming-
birds were also seen visiting Brazil nut flowers. Butterflies use
to land on the flower and insert their long proboscis to collect
nectar at any time of the day. Hawk moths were only present
early in the morning, around 0430 HR. They hovered in front
of the flowers and introduced their proboscis through the
petals to collect nectar. Hummingbirds showed no preference
for time of the day, visiting flowers at any time and also
hovered in front of the flowers to introduce their beak and
drink nectar.
4. Discussion
Results showed that B. excelsa did not set any fruit in the
restricted and hand self-pollination treatments suggesting
that this species cannot bear fruits from pollen grains origi-
nated from the same flower and requires biotic pollinators to
transfer pollen grains between flowers. According to Moritz
and M¨
uller et al. [6,10] the Brazil nut tree does not set from
self-fertilization because this mating system led to less than
the 85% ovule fertilization necessary for fruit set. However,
the geitonogamy treatment produced over 3% of fruit set
indicating that the Brazil nut tree can set fruits when pollen
grains are transferred between flowers of the same plant.
Also, results of the geitonogamy treatment were similar to
the open pollination treatment signifying that the pollination
achieved in this commercial plantation could be accounted to
geitonogamy. These findings, associated to the much greater
fruit set following hand cross pollination indicates that the
Brazil nut tree is an allogamous species, in accordance to
other authors [6,8–10].
Our results may explain why the individual plant produc-
tion is much higher in natural clusters of few Brazil nut trees
in the forest than in plantations with hundred of trees. In
the natural environment, with much fewer flowers to visit,
pollinatorsmaybeforcedtomovebetweentreesandrevisit
flowers in a much more frequent fashion than when they face
a seemly unlimited number of blooming trees.
Although many species visit Brazil nut flowers, only
some bee species showed foraging behavior compatible to
potential pollinators of this tree. While bees were numerous
and concentrated their visits to the morning shift, when
flowers presented fresh pollen and were more receptive
[10], butterflies, and hummingbirds visited inflorescences at
any time of the day, in an inconstant pattern and in low
numbers. Hawk moths, however, visited flowers in the dawn,
close to the sunrise, but were also scarce. Besides that, the
great majority of bee species entered and moved inside the
flower increasing the chance to transfer pollen from their
bodies to the stigmas, while butterflies, hawk moths, and
hummingbirds remained outside the flower and introduced
a much smaller portion of their bodies, proboscis for the
Lepidoptera and beak to the bird, being less likely to deliver
pollen to the stigmas. This behavior, in association to the
reduced number of individuals, erratic foraging activities,
and time of flower visit, suggests that these groups of floral
visitors play little or no role in the pollination of B. excelsa.
On the contrary, the foraging behavior of most bee species
indicates that they can be effective pollinators of Brazil nut
flowers, in accordance with the suggestions of Prance and
Mori [15], Mau´
es and Oliveira [20], Mau´
es [7], Zuidema
[13], and Argolo and Wadt [16].
However, some bee species could not enter the flower or
did not show a behavior suggestive of relevant pollinators
for Brazil nut. The behavior of Epicharis conica, Ep. zonata,
Megachile sp., and Centris ferruginea approaching the flower
facing in and leaving it facing out after turning its body
inside the flower can contribute to considerable deposition
on the stigma of the flower’s self-pollen (self-pollination),
showed here to produce no fruits. It may happen because
the bee leaving the flower facing out can touch the stigma
with the back of its thorax, where the pollen has just been
placed by the anthers, resulting, at the best, in a mixture of
the pollen bees carried from previously visited flowers with
that presently visited being deposited on the stigma. In such a
situation, these bee species would not be efficient pollinators
of Brazil nut flowers because B. excelsa is a predominantly
allogamous species [6,7]. Also, Apis mellifera, Melipona lat-
eralis,andFrieseomelitta longipes did not manage to enter the
flowers and could not pollinate them. Besides that, F. longipes
sometimes prevented flowers to be visited by legitimate
pollinators chasing them away for attempting to rob pollen
from their corbicula. Although this specific behavior had not
been reported before, Santos and Absy [17] showed that the
presence of other insects on the flowers can make some floral
visitors, presumably pollinators, to avoid these flowers.
Despite potential pollinators, the rare visits of Eulaema
cingulata,Epicharis umbraculata,Centris carrikeri,andEu-
friesea purpurata to Brazil nut flowers suggest that these
species contribute little to the pollination of B. excelsa.But
their presence in the trees may explain why Zuidema [13]
pointed out euglossine bees as likely pollinators of B. excelsa,
although E. umbraculata,C. carrikeri,andE. purpurata had
never before been reported as floral visitors of Brazil nut
flowers and E. cingulata only once in the study by Mau´
es [7].
Although bees of the genus Bombus had been suggested
as the main pollinators of B. excelsa [10,12], in the present
8Psyche
studyonlyoneBombus species, B. transversalis, visited the
Brazil nut flowers. Nevertheless, these visits were limited to
the onset of the blooming season. Therefore, it is likely that
the genus Bombus does not consist in a relevant taxon for the
pollination of B. excelsa in the area studied here. Similarly,
Epicharis flava and Centris americana were not abundant and
were selective in relation to the blooming stage and probably
are not among the main pollinators of Brazil nut flowers.
Bee species like Centris denudans,Eulaema meriana,Eu-
friesea flaviventris,Xylocopa frontalis,andEulaema mocsaryi
were frequent in the area during most of the blooming season
and showed body size and flower handling adequate to pol-
linate B. excelsa flowers. However, due to the abundance and
foraging behavior in the trees, we identified Eulaema moc-
saryi and Xylocopa frontalis as the most relevant pollinators
of cultivated B. excelsa in Central Amazonia. It is important
to stress that, although these two bee species are the ones
that potentially most contribute to Brazil nut pollination
under the conditions found in this study, the pollination
level achieved in the plantation is the sum of the pollination
performed by each bee species that constitute that guild of
pollinators, including those species that contributed less to
the process [21,22].
Many of the bee species presented in this study as floral
visitors and potential pollinators of Brazil nut are widespread
in the Amazon region, and some of them also occur in other
Brazilian ecosystems [23–25]. Some of these bee species were
also reported in the literature interacting with other plant
species and constitute important floral visitors or even
pollinators. Eulaema cingulata is a pollinator to Ischnosiphon
gracilis (Rudge) Koern (Marantaceae) and floral visitor of
Solanum stramonifolium Jacq. (Solanaceae) in a fragment
of the Atlantic Forest in NE, Brazil [23,24]. Vilhena and
Augusto [25] identified Ep. flava as an important floral visitor
of Malpighia emarginata in a cerrado area of Central Brazil.
In the Amazon, studies carried out in the same area of
this work on the floral biology of Bellucia grossularioides
(Melastomataceae) and floral visitors of Bixa orellana (Bixac-
eae) reported El. mocsaryi and X. frontalis as the main visitors
of these plant species [26,27]. However, only recently Santos
[17] produced the first report suggesting E. mocsaryi as an
important floral visitor and potential pollinator of B. excelsa.
Males of Eulaema meriana were observed in the present work
visiting flowers of Brazil nut to feed on nectar. According to
Williams and Whitten [28], these male bees are pollinators
of Catasetum tricomis (Orchidaceae), suggesting some level
of interdependence among these three species because the
orchid provides only essences for the male bees of Eulaema
meriana attract their conspecific females, but the pollen and
nectar necessary for the bee survival and reproduction got to
be obtained from other plant species, like the Brazil nut.
These observations support the claim of Kremen et al.
[29,30] that conserving the native vegetation on the sur-
rounding of cultivated areas is essential to keep stable pop-
ulations of pollinators, such as the bees of the present study,
for providing food, nesting, and other resources indispens-
able for their survival. The lack of effective pollinators in
numbers adequate to pollinate the large number of flowers
present in commercial plantations of Brazil nut can be a
cause for the low tree productivity observed in these areas.
Among all species identified as potential pollinators of B.
excelsa in this study only X. frontalis have been reared in
rational nest boxes and tentatively managed for pollination
of passion fruit (Passiflora edulis Sims. f. flavicarpa Deg.) in
NE Brazil [31]. Investigations on the possibility of rearing
and managing X. frontalis and other species identified here
for pollination of B. excelsa are needed.
Acknowledgments
The authors thank three anonymous referees for their
substantial contribution to this paper: Dr. Giorgio Venturieri
from Embrapa Amazˆ
onia Oriental (CPATU) Bel´
em, for his
support and relevant information; Agropecu´
aria Aruan˜
afor
permission to carry out this study in its farm, logistic
support, and use of its facilities; CAPES/Brazil for the M.Sc.
scholarship to M.C. Cavalcante and CNPq, Bras´
ılia-Brazil for
the research fellowship to B.M. Freitas (proc. No. 305062/
2007-7).
References
[1] C. A. Neves, “A Castanheira do Par´
a,” Revista de Agricultura
(Piracicaba), vol. 13, no. 10–12, pp. 463–476, 1938.
[2] A. A. Loureiro and M. F. Silva, Cat´
alogo das Madeiras da
Amaz ˆ
onia, vol. 1, SUDAM, Bel´
em—Par´
a, Brazil, 1968.
[3] A. P. Ara ´
ujo, S. Jordy Filho, and Fonseca W. N., “A vegetac¸˜
ao
da Amazˆ
onia brasileira,” in Simp´
osio do Tr´
opico ´
Umido, I,
Bel´
em. 1984, vol. 2, pp. 135–144, EMBRAPA/CPATU. 6 v.
(EMBRAPA-CPATU, Documentos, 36), Bel´
em—Par´
a, Brazil,
1986.
[4] P. Shanley, M. Cymerys, and J. Galv˜
ao, Frut´
ıferas da mata na
vida amazˆ
onica, Imazon, Bel´
em—Par´
a, Brazil, 1998.
[5] IBGE, “Produc¸˜
ao da extrac¸˜
ao vegetal e da silvicultura,” 2009.
[6] A. Moritz, Estudos Biol´
ogicos da Florac¸˜
ao e Frutificac¸˜
ao da
Castanha-do-Brasil, Embrapa Amazˆ
onia Oriental, Bel´
em—
Par ´
a, Brazil, 1984, (Embrapa Amazˆ
onia Oriental, Documen-
tos, 29).
[7] M. M. Mau´
es, “Reproductive phenology and pollination of
the brazil nut tree (Bertholletia excelsa Humb.& Bonpl.) in
Eastern Amazˆ
onia,” in Pollinating Bees—The Conservation
Link Between Agriculture and Nature, P. Kevan and I. Fonseca,
Eds., pp. 245–254, Minist´
erio do Meio Ambiente, Bras´
ılia,
Brazil, 2002.
[8] M.C.Cavalcante,Visitantes Florais e Polinizac¸˜
ao da Castanha-
do-Brasil (Bertholletia excelsa h. & b.) em Cultivo na Amazˆ
onia
Central, Dissertac¸˜
ao de mestrado, Universidade Federal do
Cear´
a, Fortaleza—Cear´
a, Brazil, 2008.
[9] B. W. Nelson, M. L. Absy, E. M. Barbosa, and G. T. Prance,
“Observation on flower visitors to Bertholletia excelsa H.B.K
and Couratari tenuicarpa A.C. Sm. (Lecythidaceae),” Acta
Amazonica, vol. 15, pp. 225–234, 1985.
[10] C. H. M ¨
uller, I. A. Rodrigues, A. A. M¨
uller, and N. R. M.
M¨
uller, Castanha-do-Brasil, Resultados de Pesquisa,no.2,
EMBRAPA-CPATU. Miscel´
anea, Bel´
em—Par´
a, Brazil, 1980.
[11] S. A. Mori, G. Prance, and A. B. Bolten, “Additional notes on
the floral biology of Neotropical Lecythidaceae,” Brittonia, vol.
30, no. 2, pp. 113–130, 1978.
[12] E. Pinheiro and M. Albuquerque, “Castanha-do-par´
a,” in
BRASIL, pp. 224–233, Minist´
erio da Agricultura. Livro Anual
da Agricultura, Bras´
ılia, Brazil, 1968.
Psyche 9
[13] P. A. Zuidema, Demography and Management of the Brazil Nut
Trees (Bertholletia excelsa), PROMAB, Riberalta, Bolivia, 2003,
Scientific series 6.
[14] F. Cornejo, Historia Natural de Casta˜
na (Bertholletia excelsa:
Lecythidaceae) y Tecnicas Para su Manejo,Asociaci
´
on para
Conservaci ´
on de la Cuenca Amazonica, Lima, Peru; The Am-
azon Conservation Association, Washington, DC, USA, 2003.
[15] G. T. Prance and S. A. Mori, “Lecythidaceae,” Flora Neotropica,
vol. 21, no. 1, pp. 1–270, 1979.
[16] V. Argolo and L. H. O. Wadt, “Abelhas visitantes de flores de
Bertholletia excelsa em ´
area de plantio e floresta nativa—Rio
Branco Acre,” in VI Congresso De Ecologia do Brasil, Fortaleza,
Editora da Universidade Federal do Cear´
a, Fortaleza—Cear´
a,
Brazil, 2003, Anais de trabalhos completos.
[17] C. F. Santos and M. L. Absy, “Polinizadores de Bertholletia
excelsa (Lecythidales: Lecythidaceae): interac¸˜
oes com abelhas
sem ferr˜
ao (Apidae: Meliponini) e nicho tr ´
ofico,” Neotropical
Entomology, vol. 39, no. 6, pp. 854–861, 2010.
[18] H. Lundberg, “Effects of weather on foraging-flights of bum-
blebees (Hymenoptera, Apidae) in a subalpine/alpine area,”
Holarctic Ecology, vol. 3, no. 2, pp. 104–110, 1980.
[19] L. Bla ˇ
zyte- ˇ
Cereˇ
skiene, G. Vaitkeviˇ
cien˙
e, S. Venskutonyt ˙
e, and
V. B ¯
uda, “Honey bee foraging in spring oilseed rape crops
under high ambient temperature conditions,” Zemdirbyste,
vol. 97, no. 1, pp. 61–70, 2010.
[20] M. M. Mau´
es and F. C. Oliveira, “Ecologia da polinizac¸˜
ao
da castanheira-do-brasil (Bertholletia excelsa) no estado do
Par ´
a,” in Proceedings of the Congresso de Ecologia do Brasil Ma-
nejo de Ecossistemas e Mudanc¸as Globais: Resumos, p. 539,
Universidade de Bras´
ılia, Departamento de Ecologia, Bras´
ılia,
Brazil, 1996.
[21] D. W. Schemske and C. C. Horvitz, “Variation among floral
visitors in pollination ability: a precondition for mutualism
specialization,” Science, vol. 225, no. 4661, pp. 519–521, 1984.
[22] M. W. Pettersson, “Pollination by a guild of fluctuating moth
populations: option for unspecialization in Silene vulgaris,”
Journal of Ecology, vol. 79, no. 3, pp. 591–604, 1991.
[23] E. L. S. Bezerra and I. C. Machado, “Biologia floral e sistema de
polinizac¸˜
ao de Solanum stramonifolium Jacq. (Solanaceae)em
remanescente de Mata Atlˆ
antica, Pernambuco,” Acta Botˆ
anica
Brasileira, vol. 17, no. 2, pp. 247–257, 2003.
[24] A. V. Leite and I. C. Machado, “Fenologia reprodutiva, biologia
floral e polinizadores de duas esp´
ecies simp´
atricas de Maran-
taceae em um fragmento de Floresta Atlˆ
antica, Nordeste do
Brasil,” Revista Brasileira de Botˆ
anica, vol. 30, no. 2, pp. 221–
231, 2007.
[25] A. M. G. F. Vilhena and S. C. Augusto, “Polinizadores da
aceroleira Malpighia emarginata DC (malpighiaceae) em ´
area
decerradonotri
ˆ
angulo mineiro,” Bioscience Journal, Uber-
lˆ
andia, vol. 23, supplement 1, pp. 14–23, 2007.
[26] C. Renner, Biologia Reprodutiva de Bellucia grossularoides
(Melastomataceae) na Regi˜
ao de Manaus,TesedeDoutorado,
Instituto Nacional de Pesquisas da Amazˆ
onia/Fundac¸˜
ao Uni-
versidade do Amazonas, Manaus—Amazonas, Brazil, 1986.
[27] F. F. Santiago, Aspectos da Biologia Reprodutiva de Bixa orellana
L. (Urucum) na Amazˆ
onia Central, Dissertac¸˜
ao de Mestrado,
Instituto Nacional de Pesquisas da Amazˆ
onia/Fundac¸˜
ao Uni-
versidade do Amazonas, Manaus—Amazonas, Brazil, 1994.
[28] N. H. Williams and M. W. Whitten, “Orchid floral fragrances
and male euglossine bees: methods and advances in the last
sesquidecade,” Biological Bulletin, vol. 164, pp. 355–395, 1983.
[29] C.Kremen,N.M.Williams,andR.W.Thorp,“Croppollina-
tion from native bees at risk from agricultural intensification,”
Proceedings of the National Academy of Sciences of the United
States of America, vol. 99, no. 26, pp. 16812–16816, 2002.
[30] C.Kremen,N.M.Williams,R.L.Bugg,J.P.Fay,andR.W.
Thorp, “The area requirements of an ecosystem service: crop
pollination by native bee communities in California,” Ecology
Letters, vol. 7, no. 11, pp. 1109–1119, 2004.
[31] B. M. Freitas and J. H. Oliveira-Filho, “Ninhos racionais para
mamangava (Xylocopa frontalis) na polinizac¸˜
ao do maracuj´
a-
amarelo (Passiflora edulis),” Ciˆ
encia Rural,vol.33,no.6,pp.
1135–1139, 2003.
