Apidologie 38 (2006) 1–12 1c© INRA/DIB-AGIB/ EDP Sciences, 2006DOI: 10.1051/apido:2006051

Original article

Botanical sources and sugar concentration of the nectarcollected by two stingless bee species in a tropical African

rain forest*

Robert Ka,b

a Faculty of Biology, Department of Behavioural Biology, Tropical Bee Research Unit PO Box 80.086, 3508 TBUtrecht, The Netherlands

b Permanent Address: Department of Forest Biology and Ecosystems Management, Faculty of Forestry andNature Conservation, Makerere University, PO Box 7062, Kampala, Uganda

Received 3 January 2005 – Revised 15 October 2005 – Accepted 10 August 2006

Abstract – Nectar foraging by species of two stingless bees was studied in an African tropical rain forest.Both species Hypotrigona gribodoi (2–3 mm) and Meliponula ferruginea (6 mm) collected nectar with awide range of sugar concentration (H. gribodoi: 14.2–67.4%; M. ferruginea: 9.1–63.4%). H. gribodoi col-lected nectar of higher sugar concentration than M. ferruginea. Factors that influenced sugar concentrationof collected nectar included botanic origin of the nectar, bee species, bee colonies, month of year, time ofday and the local environment. Sugar concentration increased gradually from 0700 h to a peak at 1300 h anddeclined thereafter. H. gribodoi collected nectar from more diverse plant species than M. ferruginea. Sugarconcentration for both bee species was higher during the dry season than the rainy season. Although theabove factors may explain part of the variation in the sugar concentration of nectar, additional explanationslie in the behavioural differences among the bee species.

nectar concentration / nectar sources / Calliandra calothyrsus / foraging behaviour / Budongo ForestReserve

1. INTRODUCTION

Several species of stingless bees can oc-cur sympatrically and their diets partly over-lap permitting access to the same resourcesby many colonies (Sommeijer et al., 1983;Roubik et al., 1986; Kleinert-Giovanini andImperatriz-Fonseca, 1987; Ramalho et al.,1989; Ramalho, 1990; Roubik et al., 1995; Ka-jobe and Roubik, 2006). The overlap in dietcan lead to competitive interactions (Hubbeland Johnson, 1977, 1978; Roubik and Aluja,1983; Roubik, 1989). The competition for nec-tar can result in the evolution of not onlymorphological adaptations of both plants andanimals but also in a wide range of nectar

Corresponding author: R. Kajobe,kajobe@forest.mak.ac.ug* Manuscript editor: B. Vassière and J. Pierre

qualities and quantities presented by plants(Roubik, 1989; Biesmeijer et al., 1999b).Flower visitors can differentiate between theamount of nectar in flowers, its sugar concen-tration, and its composition in sugars, aminoacids and lipids (Baker and Baker, 1975, 1983;Opler, 1983; Roubik and Buchmann, 1984;Roubik et al., 1995; Biesmeijer and Ermers,1999).

Honey bees have a preference mainly forthose plants that provide a better reward, thatis the largest amount of pollen, or nectar, orboth, those having more concentrated nectar,and those that have not been foraged on pre-viously by other bees (Roubik, 1989). Fac-tors that cause variation in harvested nectarsugar concentration include the botanical ori-gin of nectar, local climate, as well as be-havioural differences among bee species and

2 R. Kajobe

Figure 1. Location of theBudongo Forest Reserve(DRC = Democratic Republicof Congo).

among colonies within the same species (Bies-meijer et al., 1999b). These authors concludedthat the variation of climatological conditionsbetween patches might lead to heterogeneousnectar rewards.

Spatial and temporal niche differentiationbetween sympatric stingless bee species canarise by differences in bee morphology, for ex-ample with body size, colour and mouth parts(Biesmeijer et al., 1999a). Resource partition-ing can also be the result of behavioural dif-ferences among bee foragers, for example thetiming of the foraging activity, the ability torecruit, the intrinsic collecting behaviour, ag-gressiveness, and the food selection (Frankieet al., 1976; Villanueva, 1994). Temporal re-source partitioning was described by Frankieet al. (1976) who found that the species com-position of bee fauna foraging on a particularflower patch changed over the course of a day,thus suggesting that differences in the timingof foraging may help to avoid direct compe-tition among species at flowers. In that study,they noted a tendency for larger bees to for-age earlier in the day compared to the smallerspecies.

Studies are very limited on the floral originof nectar collected by African stingless bees.There has been no major study to fully explainthe nectar foraging ecology of social bees inAfrican tropical rain forests. The patterns andecological impact of floral choice and foragingbehaviour of social bees in this environmentare largely unknown. This study was there-fore aimed at answering the following ques-tions: what are the important nectar sources ofstingless bees in the Budongo forest reserve?What is their diversity? What factors influencethe sugar concentration of collected nectar, inparticular is there an effect of the bee specieson the concentration of the nectar collected? Acomparison between honey bees and stinglessbees was made in an effort to understand theecology of their nectar foraging behaviour.

2. MATERIALS AND METHODS

2.1. Study site

The main study was conducted in the BudongoForest Reserve (BFR), which is a tropical rain for-est located in mid-western Uganda. It is an outlier

Nectar sources and sugar concentration for African stingless bees 3

of the great Ituri forest of the Democratic Repub-lic of Congo (Fig. 1). It covers an area of 825 km2,making it Uganda’s biggest forest reserve (Hamil-ton, 1984); 53% of it is continuous forest coverwhile the remaining comprises grassland commu-nities. It is located east of the Western rift valley es-carpment and Lake Albert between 1◦35’ and 1◦55’N, and 31◦18’ and 31◦ 42’ E (Howard, 1991). Theforest lies at an average altitude of 1100 m (mini-mum 750 m and maximum 1250 m) above sea level(Eggeling, 1947).

2.2. Bee colonies

We used four natural nests of two stingless beespecies, two colonies of Hypotrigona gribodoi andtwo of Meliponula ferruginea. These two specieswere selected because they represent the two gen-era of stingless bees that have been identified inUganda. The colonies were located at most 80 mapart in various tree cavities close to the border ofthe forest. The stingless bees were identified basedon Eardley (2004) and their identity was confirmedby David W. Roubik at Smithsonian Tropical Re-search Institute. Measurement of the bee sizes (drybody length) was recorded in the Royal Museumfor Central Africa, Tervuren, Belgium, using an oc-ular micrometer on a stereoscopic microscope. Thesmaller bee species, H. gribodoi (2–3 mm) is palein colour. Two morphs of M. ferruginea (6 mm),one black and the other brown have been identifiedin Uganda. They were previously called Meliponulaerythra, but are now regarded as species synonyms(Eardley, 2004). The brown form is the one that ex-ists in Budongo forest reserve and was thereforeused in this study.

2.3. Sampling and analysis of nectarcollected by H. gribodoi andM. ferruginea

From March to August 2002, nectar sampleswere taken once a week from 10 returning foragersper colony, five times a day at 0600–0700 h, 0900–1000 h, 1100–1200 h, 1300–1400 h and 1500–1600 h local time. The methods used were similar tothose used by Biesmeijer et al. (1999b). At the nestentrances, returning nectar foragers were caughtwith a net. Nectar foragers carried pollen of dif-ferent colours and on different parts of their bodies(but did not carry pollen loads in their corbiculae).

Nectar was collected from foragers by forcing thebees to expel their crop load into a 20 µL capillarytube for the small bee, H. gribodoi and 50 µL forM. ferruginea. In the case of H. gribodoi, samplesfrom 2–3 bees were occasionally combined to ob-tain more reliable sugar concentration values. Con-centration of sugar (actually total dissolved solids)in the nectar was measured with a hand refractome-ter corrected for ambient temperature (results ex-pressed in g per 100 g of solution). Bees carryingloads with less than 5% in sugar concentration wereignored because they probably consisted mainly ofwater foragers (Roubik and Buchmann, 1984). Dataon ambient temperature was taken at each time ofnectar collection.

2.4. Identification of nectar and pollensources

Nectar foragers carried pollen of differentcolours on different parts of their bodies. Thesepollen grains adhering to the body of the beeswere sampled by trapping the bee with glycerinejelly and we analysed this pollen to identify thenectar sources used by the bees. We obtained awide variety of samples by selectively samplingforagers from each visually distinguishable pollencolour type. Such a sampling method can causea slight overestimation of the diversity in nectardiet (Biesmeijer et al., 1999b). Several representa-tives of each colour of pollen were acetolysed andmounted in glycerine jelly for microscopic exam-ination (Sawyer, 1981). The colour and amount ofthe pollen was recorded. Most slides contained onlyone type of pollen and slides that contained morethan one type were not included in our analysis.The different pollen types were identified by mak-ing use of our reference collection and by consult-ing Hamilton (1972). A nectar source was referredto as a major source if it contributed at least 30% ofthe samples from one colony during a day of obser-vation. All other sources were referred to as minorsources.

2.5. Nectar collection from Calliandracalothyrsus by stingless bees andApis mellifera

This second study was undertaken to providebackground data on the possible changes in the con-centration of the crop content of the stingless bees

4 R. Kajobe

Table I. Nectar concentration collected by foragers from four colonies of two stingless bee species upontheir return to the nest.

Colony number Number of samples analyzed Minimum Maximum Mean Standard deviationNectar concentration (g solute / 100 g solution)

Melipona ferruginea1 260 9.1 63.4 42.2 1.722 260 12.5 63.2 39.8 1.62

Hypotrigona gribodoi1 313 14.2 67.4 43.7 1.562 313 15.1 63.4 43.2 1.43

between the flowers they foraged upon and their re-turn to the nest. Also we included honey bees in thisstudy to extend our sampling to three species of so-cial bees. Data were taken once a week from Juneto September 2002 on flower visitation and concen-tration of collected nectar from C. calothyrsus (Mi-mosaceae) by A. mellifera, H. gribodoi and M. fer-ruginea in Nyabyeya Forestry College agroforestrydemonstration farm. The farm was planted withabout 5 ha of C. calothyrsus, a shrub that reached aheight of 4–6 m and flowers throughout the year andis a primary source of nectar for bees. Most flowersopen at 1600 h for a single night. Foragers were cap-tured immediately after imbibing nectar for an un-interrupted period. Nectar samples were taken oncea week from 2–5 foragers of each bee species fivetimes a day at 0600–0700 h, 0700–0800 h, 1600–1700 h, 1700–1800 h and 1800–1900 h. These timeswere selected in relation to the observed periodof anthesis of C. calothyrsus in the agro-forestryfarm. The same methods, as previously describedin 2.3 and 2.4, were used to extract the crop con-tent from the bees and identify the nectar sourcesrespectively. There was little chance the nectar thussampled came from other flower species or colonystores, since floral constancy of stingless bees isquite high (Slaa et al., 1998). The time taken tovisit one flower and five flowers in succession wasrecorded for some selected individual bees to deter-mine the handling time and foraging speed for thethree bee species gathering nectar on C. calothyr-sus.

2.6. Data analysis

The main dependent variable in this study wasthe sugar concentration of the crop content ex-pressed as gram (g) of solute per 100 g of solu-tion. These percentages ranged from 9.1 and 67.4%,and so the data was analyzed as proportion after

an Arcsine Square root Transformation (Sokal andRohlf, 1995). To report the means, the results wereconverted back to percentages. A nested ANOVAwas performed to determine the factors influencingsugar concentration of collected nectar. The modelused was:

Nectar concentration = Bee species + Colony(bee species) + Month + Time of day + Month ×Time of day +Bee species ×Month +Bee species ×Time of day + Bee species × Month × Time ofday Colony (bee species) × Month + Colony (beespecies) × Time of day + Colony (bee species) ×Month × Time of day + error. Bee species, monthand time of day were treated as fixed factors whilecolony was considered random. Pearson correla-tion coefficients were calculated to determine therelationship between the temperature and the con-centration of collected nectar. The average diver-sity of pollen types per nest was calculated usinga Shannon-Weaver index of diversity: H’ik = –Σ j

[{Nijk/ Σ j (Nijk)} ln {Nijk/ Σ j (Nijk)}] in which H’ik

is Shannon-Weaver’s index of colony i, over onemonth k and N is the number of pollen types. In-dices were calculated for each colony per monthand thereafter averaged. Kruskal-Wallis tests werecarried out to compare the nectar handling time andforaging speed on Calliandra calothyrsus.

3. RESULTS

3.1. Sugar concentration of collectednectar

The two stingless bee species H. gri-bodoi and M. ferruginea collected nectarwith a wide range of sugar concentrations(H. gribodoi: 14.2–67.4%; M. ferruginea: 9.1–63.4%; Tab. I). Overall, H. gribodoi collected

Nectar sources and sugar concentration for African stingless bees 5

Figure 2. Compared sugar concentration of nectar collected by the foragers of two species of stingless beesover 6 months in the Budongo Forest (March–August 2002).

Figure 3. Effect of the time of day (local time) on the sugar concentration of nectar collected by the foragersof two species of stingless bees polled together over 6 months in the Budongo Forest (March–August 2002).

nectar with significantly higher sugar concen-tration compared to M. ferruginea (Fig. 2)across each of the daily sampling periods(Fig. 3). Sugar concentration increased grad-ually from 0700 h in the morning to a peak at1300 h in the afternoon and there after it de-clined.

3.2. Main factors influencing sugarconcentration of collected nectar

The species of bee, colony within species,month of year and the time of day when thenectar was collected all had a significant ef-fect on the sugar concentration of collectednectar (Tab. II). There was significant differ-ence in monthly sugar concentration for the

two bee species (Tab. II). The mean nectar con-centration was higher during the dry season(June: 48.9 ± 10.8 and July: 47.8 ± 11.0) com-pared to the rainy season (March: 40.2 ± 10.0and April: 32.5 ± 9.9). There was a positivecorrelation between the concentration of col-lected nectar and temperature (M. ferruginea:2-tailed, r = 0.846, N = 260, and H. gribodoi:2 tailed, r = 0.559, N = 313, P < 0.0001 forboth species).

3.3. Nectar and pollen sources

By means of palynological analysis, 33 outof the 52 nectar sources could be identified tospecies level. Overall, more plants were ma-jor nectar sources for H. gribodoi than for

6 R. Kajobe

Table II. Nested ANOVA with the effect of bee species, colonies nested within species, month of year andtime of day on the sugar concentration of collected nectar. The blank spaces correspond to non-significantresults.

Source of variation df SS MS Fcal P valuesBee species 1 0.1056 0.1056 88.00 **Colonies (species) 2 0.0223 0.0111 9.25 *Month 5 4.1933 0.2097 174.00 **Time of day 4 12.6540 0.1318 109.83 **Month × Time of day 20Bee species ×Month 5Bee species × Time of day 4Bee species ×Month × Time of day 20Colony(species) ×Month 10Colony(species) × Time of day 8Colony(species) ×Month × Time of day 40Error 1026 1.2343 0.0012Total 1145 18.2095

Table III. Important nectar sources for two species of stingless bees in the Budongo forest over six monthsof observation (March–August 2002). Shading indicates the periods of foraging on the given source.

Plant species Family March April May June July AugustMeliponula ferruginea

Acacia sp. FabaceaeAlbizia sp. FabaceaeCalliandra calothyrsus FabaceaeCoffea sp. RubiaceaeEucalyptus sp. MyrtaceaeMusa sp. Musaceae

Hypotrigona gribodoiAcacia sp. FabaceaeAlbizia sp. FabaceaeBidens sp. AsteraceaeCalliandra calothyrsus FabaceaeCombretum sp. CombretaceaeEucalyptus sp. MyrtaceaeMusa sp. MusaceaeType 7 UnidentifiedType 14 AsteraceaeType 23 UnidentifiedSyzigium sp. MyrtaceaeVernonia sp. Asteraceae

M. ferruginea. The two most important fam-ilies were Fabaceae and Asteraceae. Three ofthe major species of Fabaceae (Acacia sp., Al-bizia sp.and Calliandra calothyrsus) were col-lected by both H. gribodoi and M. ferrug-inea (Tab. III). The three species of Asteraceae(Bidens sp., Vernonia sp and type 14) were ex-clusively collected by H. gribodoi. Eucalyp-

tus sp. being one of the two major species ofMyrtaceae was collected by both stingless beespecies. Combretaceae, Musaceae and Rubi-aceae accounted for one major source each.The families of two major types (Type 7 andType 23) could not be determined. Callian-dra calothyrsus was used throughout the ob-servation period by both stingless bee species.

Nectar sources and sugar concentration for African stingless bees 7

Table IV. Number of nectar sources and Shannon-Weaver’s index of diversity for Hypotrigona gri-bodoi and Melipona ferruginea given as averageover the 6 months of sampling.

Colony Total number of plant Diversity index (H’)species foraged upon

Meliponula ferruginea1 18 1.782 13 1.67

Hypotrigona gribodoi1 24 2.192 29 2.23

During the study period, the two stingless beespecies collected nectar from 52 different plantspecies, 13 of which were shared. M. ferrug-inea used 27 plant species (14 exclusively) andH. gribodoi used 38 plant species (25 exclu-sively). H. gribodoi collected nectar from awider range of plants species than M. ferrug-inea. The average Shannon-Weiner index ofdiversity (H’) was higher for H. gribodoi thanfor M. ferruginea (Tab. IV).

3.4. Effects of nectar source taxa onsugar concentration

The two stingless bee species shared 13nectar sources out of which H. gribodoi col-lected 8 of higher sugar concentration thanM. ferruginea (Tab. V). H. gribodoi collecteda significantly lower sugar concentration fromthe plant species it shared with M. ferrugineathan from sources that it used exclusively (t =2.69; P = 0.015; shared sources: mean = 43;N = 13; exclusive sources: mean = 51; N = 6;Tab. V). Two of the nectar sources exclusivelyused by M. ferruginea had the lowest concen-trations (Type 11 with 24.5% and Type 20 with29.2%).

3.5. Nectar collection from Calliandracalothyrsus by stingless bees andApis mellifera

C. calothyrsus was a major source of nec-tar for the three bee species and was visited by

a large variety of other insects. Observationsof flower visitation on C. calothyrsus showedthat A. mellifera was the most common visi-tor while M. ferruginea was also quite com-mon and H. gribodoi was quite rare. Amongthe bee species, A. mellifera crop content hadthe lowest mean sugar concentration (40.3 ±7.9) while H. gribodoi had the highest (44.9± 9.4; Tab. VI). The concentration of nectarin the crop of foragers of H. gribodoi andM. ferruginea captured at the nest entranceswas similar to that of the crop content frombees captured at flowers immediately after im-bibing the nectar for an uninterrupted period(H. gribodoi: ANOVA: F = 0.92; df = 1; N =85; P > 0.05; M. ferruginea: ANOVA: F =0.98; df = 1; N = 85; P > 0.05). The speciesof bee had a significant effect on the time takento visit one flower (handling time) and the for-aging speed. It was lowest for A. mellifera andhighest for H. gribodoi (One flower: Kruskal-Wallis test: df = 2, N = 85, P ≤ 0.0001; Fiveflowers: Kruskal-Wallis test: df = 2, N = 85,P < 0.0001; Tab. VI).

4. DISCUSSION

4.1. Influence of bee size on sugarconcentration of collected nectar

Distinctive preferences for nectar of dif-fering sugar concentrations were found forthe species. The smaller bee species H. gri-bodoi (2–3 mm) preferred sugar of signifi-cantly higher concentration than M. ferruginea(6 mm). If similar volumes are compared, arelatively highly concentrated nectar source isenergetically more profitable than or compa-rable to a nectar source of lower concentra-tion (Roubik, 1989). Since we used 20 µL cap-illary tubes to collect crop loads of H. gri-bodoi and 50 µL for M. ferruginea, it is hy-pothised that the crop load of the smaller bee,H. gribodoi was also smaller. Probably if allfactors were kept constant (e.g. weather con-ditions, distance to food source), the energysaved by taking less concentrated nectar wouldnot compensate foragers of H. gribodoi forthe calorific loss. By harvesting significantlymore concentrated nectar, it is possible that

8 R. Kajobe

Table V. Sugar concentration (g/100 g) of nectar of different botanic origin obtained from the crop ofreturning foragers of two stingless bee species in the Budongo forest (data collected between March andAugust 2002).

Plant species Hypotrigona gribodoi Melipona ferruginea t-value P valueMean N Mean N

Calliandra calothyrsus 45.4 147 41.2 127 2.24 0.026Eucalyptus sp. 48.1 129 44.3 98 3.04 0.003Albizia sp. 44.3 82 40.2 102 2.24 0.026Acacia sp. 49.1 76 43.2 97 2.44 0.016Musa sp. 40.4 56 41.3 64 -1.67 0.098Bidens sp. 46.2 45 38.4 39 3.40 0.001Type 14 (Asteraceae) 42.1 41 39.4 9 2.45 0.018Vernonia sp. 52.1 39 49.2 53 3.27 0.002Syzigium sp. 39.6 28 42.3 25 2.11 0.040Type 7 (Unidentified) 43.2 22 46.5 42 2.75 0.008Type 23 (Unidentified) 38.2 19 37.1 8 1.76 0.091Combretum sp. 37.8 17 32.3 11 3.61 0.001Coffea sp. 40.3 15 41.2 6 1.44 0.166Markhamia lutea 56.5 13Mangifera indica 48.4 12Ipomea sp. 54.4 9Senna sp. 47.1 6Carica papaya 58.2 5Citrus sp. 45.3 3Type11 (Unidentified) 24.5 5Type 20 (Unidentified) 29.2 4

Table VI. Sugar concentration of the crop content,handling time and foraging speed for three species ofbees gathering nectar on Calliandra calothyrsus (N = 85 for all cases; mean ± SD).

Bee species and body length Returning foragers Foragers caught at Time to visit Time to visit(measured on dry specimens) caught at nest entrance departure from flowers 1 flower 5 flowers

% TDS % TDS s sA. mellifera (11 mm) 40.3 ± 7.9 4.3 ± 1.6 26.4 ± 5.1M. ferruginea (6 mm) 41.9 ± 12.4 42.6 ± 6.8 5.2 ± 1.7 28.2 ± 4.5H. gribodoi (2–3 mm) 44.1 ± 11.8 44.9 ± 9.4 5.5 ± 2.0 31.3 ± 4.0

foragers of H. gribodoi are able to reduce theenergy needed to evaporate the collected nec-tar to the normal concentration of stingless beehoney which is about 70% (Vit et al., 2004).Roubik and Buchmann (1984) noted that evenif greatly rewarding resources are present, lessrewarding ones might be preferred if the over-all amount of that resource in the patch is highand/or easy to forage. However, a sample sizeof two stingless bee species with two coloniesof each is far too small to enable the drawingof general conclusions as to the importance ofthe size factor on the nectar selection processof stingless bees.

The two bee species collected nectar witha wide range of sugar concentrations suggest-ing that the concentrations used partly over-lapped between the species. The wide range ofsugar concentration can be favourable to bothbee species because it can ultimately result insurvival of these species in cases of adverseenvironmental conditions and inadequate nec-tar resources. Accessibility to nectar sourcesby the bee species may have limited influenceon the concentration of collected nectar be-cause the bee species collected nectar fromplant species in the same families and sharedsome of their major nectar sources.

Nectar sources and sugar concentration for African stingless bees 9

4.2. Effects of nectar and pollen sourceon sugar concentration of nectar

Bees must possess a mechanism thatcan discriminate between nectars of differ-ent sweetness to be able to collect the opti-mal sugar concentration. Other possible mech-anisms are distinction by taste or sensitivity tosugar concentration (Biesmeijer et al., 1999b).Variation in plant species used by bee speciesis one of the factors that explain the variationin sugar concentration of bee collected nectar.The two bee species shared 13 sources, someof which were among their major sources. Ofthe total range of nectar and pollen sourcesused by the two species, H. gribodoi tended touse plants with higher average nectar concen-tration. Moreover, the shared nectar sourceswere those that provided significantly less richnectar. Thus the botanical origin of the nectarseems to explain only part of the differences inthe sugar harvest of the two species.

4.3. Temporal influence on sugarconcentration of collected nectar

Another factor that may explain some of thevariation in sugar concentration of collectednectar is time of day. This study revealed thatthe sugar concentration of the bee collectednectar increased from early morning till 1300–1400 h and declined thereafter. This finding isin line with the prediction that higher solar ra-diation causes evaporation and results in moreconcentrated nectar in the flowers (Roubik andBuchmann, 1984; Roubik; 1989). Sugar con-centration of collected nectar varied during themonths (seasons) of data collection. Monthlychanges in flowering phenology are likely toproduce differences in nectar richness. Withmore rain and fewer sunny periods, there willbe a low concentration of bee collected nec-tar. During the dry season, the wind evaporatessome of the water in the nectar thereby makingit more concentrated.

4.4. Behavioural factors regulatingpartitioning of nectar resources

The diet diversity of H. gribodoi was higherthan that of M. ferruginea. The recruitment

strategies of the bees to food sources mayprobably explain the differences in diversity.In a related study (Kajobe and Echazarreta,2005) found that workers of M. ferrugineaexited their nests in characteristically distinctforaging bouts suggesting that the recruit-ment methods employed may be direct lead-ing or “piloting” (Esch et al., 1965; Esch,1967; Johnson, 1987). Such recruitment traitcould have led to the reduced diversity in nec-tar sources for M. ferruginea. This hypothesisneeds to be tested further because there is littledoubt that it is adaptive for a bee to occasion-ally sample flowers other than those on whichit is specialising. Such behaviour may fre-quently occur in tropical habitats where hun-dreds of flower species can be presented at onetime (Roubik and Buchmann, 1984).

4.5. Nectar collection from Calliandracalothyrsus by stingless bees andApis mellifera

Despite the fact that bees might have har-vested nectar from different flowers, there wasno significant difference between the con-centration of nectar harvested by foragers ofH. gribodoi and M. ferruginea captured at thehive entrances as compared to nectar frombees captured at flowers. This is in agree-ment with Roubik et al. (1995) who con-cluded that nectar recently stored in the cropof bees is not dehydrated. A. mellifera wasthe most common forager on C. calothyrsus.Such a situation could have arisen as a resultof contest competition (Milinski and Parker,1991) during which the stingless bees are dis-placed from the resource as a result of ag-gressive behaviour of A. mellifera. Althoughthe competing stingless bees generally sharedthe resource, M. ferruginea, was competitivelystronger than H. gribodoi and exploited theshared resource faster.

Various sizes and shapes of bee mouthpartsmay results in differences in rate at whichthey harvest nectar of differing concentrationand from different floral types. Such mor-phological differences may ultimately lead tonectar resource partitioning. Apart from thedifferences that floral structure enforces inregulating visitor type, resource partitioning

10 R. Kajobe

should reflect parameters of foraging successthat are dependent upon the physical proper-ties of nectar and upon the bees themselves(Roubik and Buchmann, 1984). Comparisonof the nectar intake rate of the three beespecies shows that the larger bees tended toharvest nectar faster. Fast foragers, which sug-gest higher nectar intake rate, collected moredilute nectar. By imbibing more dilute nectarat a faster rate, foragers of A. mellifera mayhave reached a balance between foraging en-ergetics and profit.

Based on these results, it is concluded thatthe botanic origin of the nectar, preference forcertain sugar concentration, local environmen-tal conditions, temporal influences and colonyrecruitment behaviour may be used to explainpart of the variation in sugar concentration ofthe nectar collected. Other factors (not testedin this study) that may influence sugar concen-tration of the nectar collected include distanceof flowers from a bee nest, the number andarrangement of available flowers, floral mor-phology, and the willingness of nest mates toaccept nectar from a returning forager (Roubikand Buchmann, 1984; Roubik et al., 1995).

ACKNOWLEDGEMENTS

I am grateful to the British Ecological Society(BES) and the Faculty of Forestry and Nature Con-servation, Makerere University for funding this re-search. I especially thank Dr. J.R.S. Kaboggoza andProfessor J. Obua for their support. I am grateful toDr. David W. Roubik who confirmed the identity ofthe stingless bee species. I also thank the Royal Mu-seum for Central Africa, Tervuren, in Belgium whosponsored me to study their bee collection in June2003.

Résumé – Sources botaniques et concentrationen sucres du nectar récolté par deux abeilles sansaiguillon dans une forêt pluviale tropicale afri-caine. Une étude a été menée de mars à août 2002sur la récolte de nectar par deux abeilles sans ai-guillon (Hymenoptera, Apidae) dans la réserve dela forêt de Budongo, dans le centre ouest de l’Ou-ganda, avec comme objectif de déterminer les fac-teurs qui influencent la concentration en sucres dunectar. Deux nids naturels de Hypotrigona gribodoiet deux de Meliponula ferruginea ont été utilisés.Les échantillons de nectar ont été prélevés une foispar semaine sur des 10 butineuses par colonie quirentraient au nid, cinq fois par jour. Une seconde

expérience a été menée sur Calliandra calothyr-sus pour obtenir des données sur les variations pos-sibles de la concentration du jabot de ces abeillesentre le moment où elles butinaient les fleurs et ce-lui où elles rentraient au nid. Les abeilles domes-tiques (Apis mellifera) aussi ont été incluses danscette étude. Les échantillons de nectar ont été pré-levés une fois par semaine sur 2 à 5 butineuses dechaque espèce cinq fois par jour.Les deux abeilles sans aiguillon ont récolté du nec-tar dont la concentration en sucres variait grande-ment (H. gribodoi : 14,2–67,4 % ; M. ferruginea :9,1–63,4 % ; Tab. I). La concentration du nectar ré-colté par H. gribodoi était significativement plusforte que celui récolté par M. ferruginea (Fig. 1) etce à toutes les périodes d’échantillonnage au coursde la journée (Fig. 3). La concentration en sucress’est accrue le matin pour atteindre un pic l’après-midi puis redescendre. L’espèce d’abeille, la colo-nie au sein de la même espèce, le mois de l’annéeet l’heure de l’échantillonnage ont tous eu un effetsignificatif sur la concentration en sucres du nectarrécolté (Tab. II).Sur 52 sources de nectar utilisées par les abeilles, 33ont été identifiées au niveau spécifique. H. gribodoia utilisé 38 plantes et M. ferruginea 27 et les deuxespèces en ont partagé 13 (Tab. V). C. calothyrsusa constitué une source importante pour les trois es-pèces d’abeilles, A. mellifera étant la plus fréquentesur cette plante. Le contenu du jabot d’A. melliferaavait la plus faible concentration en sucres (40,3 ±7,9), celui de H. gribodoi la plus élevée (44,9 ± 9,4 ;Tab. VI). La concentration du nectar dans le jabotdes deux abeilles sans aiguillon capturées à l’entréede leur colonie était la même que celle du jabot desabeilles capturées sur les fleurs.La large gamme de concentration en sucres peut fi-nalement aboutir à sélectionner ces espèces en casde conditions environnementales défavorables et deressources en nectar inadaptées. On a trouvé denettes différences pour des nectars en fonction deleur concentration en sucres. En récoltant un nectarplus concentré, il est possible que les butineuses deH. gribodoi soient capables de réduire l’énergie né-cessaire pour évaporer le nectar récolté et l’amenerà la concentration normale du “ miel ” d’abeillessans aiguillon (70 % environ). L’accroissement dela concentration en sucres du matin jusqu’à l’après-midi est conforme à la prédiction selon laquelle laradiation solaire plus forte provoque de l’évapora-tion et aboutit à un nectar plus concentré. L’originebotanique du nectar, la préférence pour certainesconcentrations en sucres, l’environnement local etle comportement de recrutement peuvent expliqueren partie la variation de la concentration en sucresdu nectar récolté.

concentration en sucres / nectar / plante necta-rifère / comportement de butinage / abeille sansaiguillon / Apidae / Calliandra calothyrsus / Ou-ganda

Nectar sources and sugar concentration for African stingless bees 11

Zusammenfassung – Botanische Herkunft undZuckerkonzentration des Nektars, gesammelt ineinem afrikanischen Regenwald von zwei Ar-ten der Stachellosen Bienen. Das Nektarsammelnzweier Arten von Stachellosen Bienen wurde vonMärz bis August im Budonge Waldreservat im mit-telwestlichen Uganda untersucht. HauptsächlichesZiel war, die Einflussfaktoren der Zuckerkonzen-tration des gesammelten Nektars zu bestimmen.Es wurden jeweils zwei natürliche Nester von Hy-potrigona gribodoi und Meliponula ferruginea ge-nutzt. Einmal pro Woche wurden zu 5 verschie-denen Tageszeiten Nektarproben von 10 zurück-kehrenden Sammlerinnen genommen. Ein weiteresExperiment zur Bestimmung möglicher Änderun-gen der Konzentration des Honigmageninhalts zwi-schen den besammelten Blüten und der Rückkehrzum Nest wurden an der Stachellosen Biene Cal-liandra calothyrsus unternommen. In dieses Expe-riment wurden auch Honigbienen eingeschlossen.Einmal pro Woche wurden 5 mal pro Tag Nektar-proben von 2–5 Sammlerinnen jeder Bienenart ge-nommen.Der von den beiden Stachellosen Bienen gesammel-te Nektar umfasste eine weite Spanne unterschied-licher Zuckerkonzentrationen (H. gribodoi: 14,2–67,4 %; M. ferruginea: 9,1–63,4 %; Tab. I). H. gri-bodoi sammelte zu jeder der untersuchten Tageszei-ten Nektar mit signifikant höherer Zuckerkonzen-tration als M. ferruginea (Abb. 2, 3). Die Zucker-konzentration nahm vom Morgen bis zum Spitzen-wert am Nachmittag zu und nahm dann wieder ab.Die Bienenart, die Kolonie innerhalb der Spezies,der Jahresmonat und die Tageszeit, zu der Nektargesammelt wurde, hatten alle einen signifikantenEffekt auf dessen Zuckerkonzentration (Tab. II).Von den 52 besammelten Nektarquellen wurden 33bis zur Art bestimmt. H. gribodoi nutzte 38 Pflan-zenarten, M. ferruginea 27. Beide Bienenarten hat-ten 13 Nektarquellen gemeinsam (Tab. V). C. ca-lothyrsus war für drei Bienenarten die Hauptquel-le (Apis mellifera, H. gribodoi und M. ferruginea),hierbei war A. mellifera der häufigste Besucher.Der Honigblaseninhalt von A. mellifera wies dieniedrigste mittlere Zuckerkonzentration auf (40,3 ±7,9), bei H. gribodoi war sie am höchsten (44,9 ±9,4; Tab. VI). Die Nektarkonzentration in der Ho-nigblase von am Nesteingang gefangenen H. gribo-doi und M. ferruginea war ähnlich zu der von anBlüten abgefangenen Bienen.Der weite Bereich von Zuckerkonzentrationenkann ultimativ zu einer Auslese dieser Arten beischlechten Umgebungsbedingungen oder ungeeig-neten Nektarquellen führen. Es wurden klare Be-vorzugungen für Nektare unterschiedlicher Zucker-konzentrationen gefunden. Möglicherweise könnenSammlerinnen von H. gribodoi durch das Sammelnvon höher konzentriertem Nektar die zur Verdun-stung des gesammelten Nektars bis zu der für Sta-chellose Bienen normalen Konzentration von 70 %benötigte Energie reduzieren. Das Ansteigen der

Zuckerkonzentration des von den Bienen gesam-melten Nektars vom Morgen bis zum Nachmittagstimmt mit der Vorhersage überein, dass die durchdie höhere Sonneneinstrahlung verursachte Verdun-stung zu einer höheren Nektarkonzentration in denBlüten führt. Die botanische Herkunft des Nek-tars, die Bevorzugung bestimmter Zuckerkonzen-trationen, die örtliche Umgebung, zeitliche Einflüs-se auf das Rekrutierungsverhalten der Völker kön-nen einen Teil der Variation in den Zuckerkonzen-trationen des gesammelten Nektars erklären.

Nektarkonzentration / Nektarquellen / Callian-dra calothyrsus / Sammelverhalten / BudongoWaldreservat / Uganda

REFERENCES

Baker H.G., Baker I. (1975) Studies of nectar consti-tution and pollinator-plant coevolution, in: GilbertL.E., Raven P.H. (Ed.), Animal and plant coevolu-tion, Austin, Texas, USA, pp. 100–138.

Baker H.G., Baker I. (1983) Floral nectar sugar con-stituents in relation to pollinator type, in: JonesC.E., Little R.J. (Eds.) Handbook of experimentalpollination biology, Van Nostrand Reinhold, NewYork, pp. 117–141.

Biesmeijer J.C., Ermers M.C.W. (1999) Social forag-ing in stingless bees: how colonies of Meliponafasciata choose among nectar sources, Behav.Ecol. Sociobiol. 46, 129–140.

Biesmeijer J.C., Richter J.A.P., Smeets M.A.J.P.,Sommeijer M.J. (1999a) Niche differentiation innectar collecting stingless bees: the influence ofmorphology, floral choice and interference com-petition, Ecol. Entomol. 24, 1–9.

Biesmeijer J.C., Smeets M.J.A.P., Richter J.A.P.,Sommeijer M.J. (1999b) Nectar foraging by stin-gless bees in Costa Rica: botanical and climato-logical influences on sugar concentration of nectarcollected by Melipona, Apidologie 30, 43–55.

Eardley C.D. (2004) Taxonomic revision of theAfrican stingless bees (Apoidea: Apidae: Apinae:Meliponini), Afr. Plant Prot. 10, 63–96.

Eggeling W.J. (1947) Observations on the ecology ofBudongo rain forest, Uganda, J. Ecol. 44, 20–87.

Esch H., Esch I., Kerr W.E. (1965) Sound: an elementcommon to communication of stingless bees andto dances of the honey bee, Science 149, 320–321.

Esch H. (1967) The evolution of bee language, Sci.Am. 216, 97–104.

Frankie G.W., Opler P.A., Bawa K.S. (1976) Foragingbehaviour of solitary bees: Implications for out-crossing of a neotropical forest tree species, J.Ecol. 64, 1049–1057.

12 R. Kajobe

Free J.B., Pickett J.A., Ferguson A.W., SimpkinsJ.R., Williams C. (1984) Honeybee Nasonovpheromone lure, Bee World 65, 175–181.

Hamilton A.C. (1972) The interpretation of pollen di-agrams from highland Uganda, Paleoecol. Africa7, 46–149.

Hamilton A. (1984) Deforestation in Uganda, OxfordUniversity Press, Nairobi.

Howard P.C. (1991) Nature conservation inUganda’s tropical forest reserves. IUCN, Gland,Switzerland.

Hubbell S.P., Johnson L.K. (1977) Competition andnest spacing in a tropical stingless bee community,Ecology 58, 949–963.

Hubbell S.P., Johnson L.K. (1978) Comparative for-aging behaviour of six stingless bees exploiting astandardized resource, Ecology 59, 1123–1136.

Johnson L.K. (1987) Communication of food sourcelocation by the stingless bee Trigona fulviven-tris, in: Eder and Rembold (Ed.), Chemistry andBiology of Social Insects, Neotropical Bees. EdercRembold, Verlag, Peperny, Munchen.

Johnson L.K., Hubbell S.P. (1974) Aggression andcompetition among stingless bees: field studies,Ecology 55, 120–127.

Kajobe R., Echazarreta C.M. (2005) Temporal re-source partitioning and climatological influenceson colony flight and foraging of stingless bees(Apidae; Meliponini) in Ugandan tropical forests,Afr. J. Ecol. 43, 267–275.

Kajobe R., Roubik D.W. (2006) Honey-making beecolony abundance and predation by apes and hu-mans in a Uganda Forest Reserve, Biotropica 38,1–9.

Kleinert-Giovannini A., Imperatriz-Fonseca V.L.(1987) Aspects of the trophic niche of Meliponamarginata marginata, Lepeletier (Apidae,Meliponinae), Apidologie 18, 69–100.

Milinski M., Parker G.A. (1991) Competition forresources, in: Krebs J.R., Davies N.B (Eds.),Behavioural Ecology – an Evolutionary approach,3rd ed., Blackwell Scientific Publications, Oxford,UK, pp. 137–168.

Opler P.A. (1983) Nectar production in the tropicalecosystem, in: Bentley B., Elias T. (Eds.), The bi-ology of nectarines, Columbia University Press,New York, pp. 30–79.

Ramalho M. (1990) Important bee plants for stinglessbees (Melipona and Trigonini) and Africanizedhoney bees (Apis mellifera) in neotropical habi-tats: a review, Apidologie 21, 469–488.

Ramalho M., Kleinert-Giovannini A., Imperatriz-Fonseca V.L. (1989) Utilization of floral resourcesby species of Melipona (Apidae, Meliponinae):floral preferences, Apidologie 20, 185–195.

Rohlf F.J., Sokal R.R. (1995) Statistical tables, 3rd ed.,W.H. Freeman, San Francisco, CA.

Roubik D.W. (1989) Ecology and natural history oftropical bees, Cambridge University Press, NewYork.

Roubik D.W., Aluja M. (1983) Flight rangesof Melipona and Trigona in tropical forest(Hymenoptera: Apidae), J. Kans. Entomol. Soc.56, 217–222.

Roubik D.W., Buchmann S.L. (1984) Nectar selectionby Melipona and Apis mellifera (Hymenoptera:Apidae) and the ecology of nectar intake by beecolonies in a tropical forest, Oecologia 61, 1–10.

Roubik D.W., Moreno J.E., Vergara C., WittmannD. (1986) Sporadic food competition with theAfrican honey bee: projected impact on neotrop-ical social bees, J. Trop. Ecol. 2, 97–111.

Roubik D.W., Yanega D., Alluja S.M., Buchmann S.L.,Inouye D.W. (1995) On optimal nectar foragingby some tropical bees (Hymenoptera: Apidae),Apidologie 26, 197–211.

Sawyer R.W. (1981) Pollen identification for beekeep-ers, UK University Collage, Cardiff Press, Cardiff,Wales.

Slaa E.J., Cevaal A., Sommeijer M.J. (1998) Floralconstancy in Trigona stingless bees (Apidae,Meliponinae) foraging on artificial flower patches:a comparative study, J. Apic. Res. 37, 191–198.

Sommeijer M.J., De Rooy G.A., Punt W., de BruijnL.L.M. (1983) A comparative study of forag-ing behaviour and pollen resources of variousstingless bees (Hym., Meliponinae) and honey-bees (Hym., Apidae) in Trinidad, West-Indies,Apidologie 14, 205–224.

Villanueva R.G (1994) Nectar sources of European andAfricanized honey bees (Apis mellifera L.) in theYucatan Peninsula, Mexico, J. Apic. Res. 33, 44–58.

Vit P., Medina M., Enriquez M.E. (2004) Quality stan-dards for medicinal uses of Meliponinae honey inGuatemala, Mexico and Venezuela, Bee World 85,2–5

Willmer P.G., Corbet S.A. (1981) Temporal and mi-croclimatic partitioning of the floral resources ofJusticia aurea amongst a concourse of pollen vec-tors and nectar robbers, Oecologia 51, 67–78.