HAL Id: hal-00891483https://hal.archives-ouvertes.fr/hal-00891483

Submitted on 1 Jan 1998

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Protein trophallaxis and the regulation of pollenforaging by honey bees (Apis mellifera L.)

Scott Camazine, Karl Crailsheim, Norbert Hrassnigg, Gene E. Robinson,Bernhard Leonhard, Helga Kropiunigg

To cite this version:Scott Camazine, Karl Crailsheim, Norbert Hrassnigg, Gene E. Robinson, Bernhard Leonhard, etal.. Protein trophallaxis and the regulation of pollen foraging by honey bees (Apis mellifera L.).Apidologie, Springer Verlag, 1998, 29 (1-2), pp.113-126. <hal-00891483>

Original article

Protein trophallaxis and the regulation of pollenforaging by honey bees (Apis mellifera L.)

Scott Camazine’ Karl Crailsheimb Norbert HrassniggGene E. Robinson Bernhard Leonhard Helga Kropiunigg

"Department of Entomology, Penn State University,501 ASI Building, University Park, PA 16802, USAbInstitut für Zoologie an der Karl-Franzens-Universität,

Universitätsplatz 2, A-8010 Graz, AustriacDepartment of Entomology, University of Illinois, Urbana, IL 61801, USA

(Received 15 July 1997; accepted 28 October 1997)

Abstract - Pollen foragers quickly sense increases in colony pollen stores, and modify theirforaging activity appropriately. In association with these changes in foraging behavior, nurse beestransfer a larger portion of newly synthesized 14C-phenylalanine-labeled protein to the foragers.These findings support the hypothesis that trophallactic interactions between nurse bees andpollen foragers may serve as a cue apprising pollen foragers of the colony’s need for pollen.&copy; Inra/DIB/AGIB/Elsevier, Paris

pollen forager / protein trophallaxis / nurse bee

1. INTRODUCTION

1.1. Information acquisitionand colony-level regulation

Social insects precisely regulate colony-level processes critical for their survival.For example, honey bees modulate their

intensity of nectar and pollen foraging inaccordance with colony needs (Lindauer,1952; Free, 1967; Cale, 1968; Barker,1971; Free and Williams, 1971; van Laereand Martens, 1971; Al-Tikrity et al., 1972;Moeller, 1972; Seeley, 1989; Seeley et al.,1991; Fewell and Winston, 1992; Seeleyand Towne, 1992). Fire ants apportion for-aging efforts among carbohydrates, pro-

* Correspondence and reprintsTel.: (1) 814 863 1854; fax: (1) 814 865 3048; e-mail: SMC14@psu.edu

tein and lipids (Sorensen et al., 1985).They also regulate the rate of egg layingby the queen (Tschinkel, 1988). In

response to changing environmental con-ditions, honey bees carefully control thetemperature of their nest or swarm (Lin-dauer, 1954; Heinrich, 1981a, b, c) andtermites also exhibit sophisticated regu-lation of the temperature, carbon dioxidelevel and humidity of their nest (Lüscher,1961).

How do social insects accomplish thesetasks? All regulatory mechanisms requireacquisition of information about the stateof the system and the implementation ofappropriate regulatory responses. Multi-cellular organisms such as humans use’centralized’ homeostatic mechanisms to

regulate physiological variables such asbody temperature and glucose level. Bymeans of a dense network of circulatoryand neural pathways that permeate thebody, information from throughout theorganism can be acquired and processedcentrally, by the brain for example, allow-ing the execution of the appropriate behav-ioral or physiological responses.

But the mechanisms used by a singleorganism are not the same as those used bya colony of many organisms. A colony ofsocial insects may consist of thousands oreven millions of autonomous individuals.It is unlikely that an individual colonymember could acquire detailed informa-tion about the state of the entire colony.Therefore one would not expect specificindividuals to play central roles in direct-ing the activity of other colony members.In lieu of centralized mechanisms of

colony organization, social insects haveevolved efficient decentralized mecha-nisms based upon each individual’s

response to information acquired in itslocal environment. As such, to understandthe regulatory mechanisms underlyingcolony-level processes, one must deter-mine what information individuals acquireabout colony needs, and how they use that

information to modulate their behavior.This study addresses this question in thecontext of the regulation of pollen forag-ing by honey bees.

1.2. Protein balance in honeybee colonies

Protein for honey bee larval develop-ment is derived from the digestion ofpollen. Adult bees also require protein(Crailsheim, 1986, 1990a). However, notall bees consume and metabolize pollenequally well. Workers approximately 8days old are the colony’s primary pollenprocessors and distributors. They act asnurses, feeding proteinaceous hypopha-ryngeal gland secretion (jelly) to the lar-vae. They also feed jelly to other colonymembers (Crailsheim, 1990a, b, 1991).The gastrointestinal tract of these pollenprocessor bees has the highest pollen con-tent and their midguts have the highestproteolytic activity of all bees in the colony(Moritz and Crailsheim, 1987; Crailsheimet al., 1992). In addition, their hypopha-ryngeal glands are especially well devel-oped (Moritz and Crailsheim, 1987; Crail-sheim and Stolberg, 1989). Throughoutthis paper we refer to these bees inter-

changeably as nurse bees or pollen pro-cessors.

In contrast, pollen foragers themselvesconsume little pollen, have little enzymaticcapability to digest pollen, and have atro-phied hypopharyngeal glands (Moritz andCrailsheim, 1987; Crailsheim, 1990a).Thus, the pollen processors serve as thecolony’s consumers and distributors ofprotein, while the foragers act as the pollencollectors.

1.3. The regulation of pollen foraging

A previous study (Camazine, 1993)showed that pollen foragers can quickly

acquire information about the colony’sneed for pollen, and that the foragers donot need ’direct’ contact with the supple-mented pollen to detect the colony’schange of state. Instead pollen foragersare able to obtain their information about

colony pollen need ’indirectly’ from otherbees in the hive. It was suggested that alikely source of information was the pollenprocessors. As the major consumers ofpollen and distributors of protein, thesebees are in a pivotal position to integrate aswell as disseminate information concern-

ing the supply and demand for pollenwithin the colony.

This raises the question of what infor-mation the pollen foragers are using andhow they obtain information.

1.4. An hypothesis for the regulationof pollen foraging

How might pollen processors commu-nicate information about the colony’s needfor pollen? We focus on the possibilitythat trophallactic interactions betweenpollen processors and pollen foragers pro-vide automatic and reliable information

concerning the colony’s need for pollen.When the colony has ample pollen stores,pollen processors may have large amountsof protein-rich jelly available to distributeto the pollen foragers. The consumptionof sufficient jelly by the pollen foragersmay inhibit their foraging. In contrast,under conditions of pollen dearth, thepollen processors may have less proteinavailable for the foragers, thereby stimu-lating them to forage for pollen.

In this paper we address the questionof whether pollen processors provideinformation to pollen foragers concerningthe status of the colony’s proteinaceouspollen stores. In particular, we address thefollowing two questions: 1) How dochanges in the colony’s pollen stores affectthe behavior of the pollen foragers? 2)

How do changes in colony’s pollen storesaffect the trophallactic behavior and phys-iology of the pollen processors?

2. MATERIALS AND METHODS

2.1. Study sites and honeybee colonies

For each of the three experiments, we pre-pared a pair of colonies matched for adult pop-ulation; amount and age of brood; and honeyand pollen stores. The two colonies werearranged so that their entrances were 2 m apart.Each colony had two frames containing broodand honey. The contents of the third frame var-ied with each experiment, as described below.Experiment 1 was conducted in Urbana, Illi-nois, using Italian honey bees (Apis melliferaligustica) kept in observation hives with threefull-depth Langstroth frames, and internaldimensions of 78 x 46.5 x 4.5 cm. Experiments2 and 3 were conducted in Graz, Austria, usingcarnica honey bees (A. m. carnica). Thesecolonies were kept in observation hives withinternal dimensions 78 x 44 x 4 cm, containingthree 42 x 22 cm frames. The colonies wereset up 2 weeks before the start of each experi-ment.

The colonies for experiments 1, 2 and 3,contained approximately 8 500, 6 800 and7 500 bees, respectively. Population countswere made by marking off a 5 cm square gridon the glass walls of the observation hive, andcounting the number of bees in each gridsquare.

2.2. Experimental design

The experimental design for these experi-ments involved setting up two colonies in par-allel, depriving the bees of pollen for severaldays, and then supplementing one colony withpollen, while keeping the other colony pollendeprived. During the treatment period, analy-ses were made of the behavior of the pollenforagers, trophallactic interactions betweenpollen processors and pollen foragers, andphysiological changes of the pollen proces-sors.

2.2.1. Introduction of newly-emerged,marked bees

Seven days before day 1 of each experi-ment, 200 newly emerged bees were markedand placed into each colony. These bees hadbeen taken from an unrelated colony. On day 1of the experiment, these bees were 8 days old,the age which corresponds to the approximatepeak of their pollen consumption and nursingactivity (Crailsheim, 1991).

2.2.2. Pollen deprivation

For 5 days prior to the start of each exper-iment, both colonies were deprived of pollen.Each colony had approximately equal amountsof brood located on the upper and middleframes of the observation hive. A strip of queenexcluder material was placed between thesetwo upper frames and the lower frame, restrict-ing the queen to the upper portion of the hive.As a result, the two upper frames becamenearly full of brood, except for a small rim ofhoney. There was little pollen on the upperframes; at the end of the day, fewer than 20pollen cells were found, and by the next morn-ing, little, if any, pollen was seen. The lowerframe in each hive was removed at the end ofeach day (between 1700 and 1900 hours) andreplaced with an empty frame, or a frame withsome honey, depending upon whether thecolony had sufficient honey stores. Sincealmost all the pollen collected by the foragerswas deposited on the lower frame, and sincethis frame was removed each evening, thecolony was relatively starved for pollen dur-ing this preparatory period.

2.2.3. Experimental day I

On the first day of each experiment, wetreated both colonies identically: each had anempty frame in the lower position of the obser-vation hive, placed in the colony the eveningbefore as a part of the pollen deprivation pro-tocol. Starting at approximately 0800 hours,when the first foragers began to return to thehive, each returning pollen forager was iden-tified by the presence of pollen loads in hercorbiculae. The pollen foragers from each ofthe two hives were marked with a differentcolor paint, spotted on the thorax or abdomenwith a fine paintbrush. Paints were mixed usingshellac, dry pigments and sufficient 96 %

ethanol to obtain the proper consistency. Overthe next 6-8 h, every unmarked pollen foragerwas marked as she returned to the hive, and arecord was kept of the number of bees markedduring each 15-min period. We stopped mark-ing bees between 1400 and 1645 hours whenthe number of returning pollen foragers fellbelow approximately five bees per 15-minperiod. Then, the observation hives wereopened, the queen excluder screens removed,and each colony was given a treatment. Onecolony was randomly selected as the experi-mental colony, and its lower frame was

replaced with a frame approximately half-fullof pollen. (The remainder of the frame wasempty.) The other colony was designated thecontrol colony, and it was given an emptyframe. These frames remained in the colonyovernight and during the next day, experimentalday 2.

2.2.4. Injection of marked beeswith radioactive amino acid

On the evening of day 1, at approximately2100 hours (between 4 and 5 h after the exper-imental colony received supplemental pollen),50 of the marked 8-day-old bees were takenfrom each colony, injected with 1 &mu;L 14C-phenylalanine (18.4 GBq/mmol, 3.7 MBq/mL,0.033 mg/mL, from NEN), and returned totheir colony. A second batch of 12 markednurses was killed after 1.5 h and 1 &mu;L ofhemolymph was taken from each bee. Thishemolymph sample was frozen at -20 °C, andkept for amino acid analysis. The gastroin-testinal tract (GI) of these bees was alsoremoved for measurements of midgut weightand protein content.

2.2.5. Experimental day 2

In experiments 2 and 3, each colony wasopened at approximately 0700 hours, and allthe radiolabeled nurse bees were removed for

analysis of radioactive label. In addition, 12non-injected, marked nurses were killed forhemolymph amino acid analysis, and GI tractweight and protein content to compare withthe samples taken on the evening of day 1.Twelve pollen foragers and 12 switchers (beesthat were pollen foragers on day 1, but switchedto nectar foraging on day 2) were also preparedfor analysis of midgut weight and protein con-tent as they returned from foraging on the

morning of day 2. As on day 1, starting atapproximately 0800 hours, when pollen for-agers began to return to the hive, each bee thathad been marked as a pollen forager on theprevious day, was again marked with coloredpaint as she returned to the hive. The bees weremarked with a different color depending uponwhether they returned to the hive with pollen orwithout pollen. This resulted in two categoriesof bees: 1) bees that had foraged for pollen onday 1, and that continued to forage for pollenon day 2 (hereafter called ’continuing pollenforagers’), and 2) bees that had foraged forpollen on day 1, but that had not foraged forpollen on day 2. (These bees are foragers thatpresumably switched from pollen foraging onday 1 to nectar foraging on day 2, and are here-after called ’switchers’.) In addition, at the endof the day, the observation hive was opened,and all the bees were collected which had for-

aged for pollen on day 1, but had not left thehive on day 2. (These bees were easily identi-fied because they had a paint mark from havingcollected pollen on day 1, but had no paintmark indicating that they had returned to thehive on day 2.) These bees made up a third cat-egory of foragers: 3) bees that had foraged forpollen on day 1, but had not foraged at all onday 2, and are hereafter called ’quitters’.

2.2.6. Analysis of the levelsof radioactivity in the nursebees and foragers

The nurse bees and foragers were keptfrozen until analysis. They were then burned inan Packard Oxidizer where CO2 was trappedwith Carbosorb that was mixed with HionicFluor, a liquid scintillation cocktail (both fromPackard). Counts of 14C activity were carriedout with a Packard 1900CA Tri-Carb liquidscintillation counter. Results are given as DPMs(disintegrations per minute).

2.2.7. Analysis of midgut weightand protein content

Honey bees were individually dissected,and the midgut along with its contents wasremoved and weighed to the nearest 0.1 mg.The protein content of the midgut was thendetermined according to the method of Lowryet al. (1951), with bovine serum albumin (fromSigma: fraction V) as a standard. The tissueswere put into 0.5 mL of a 1 M NaOH solution

and homogenized for 30-45 s with an ultra-sonic homogenizer. Then, an additional 0.5 mLof NaOH was added. The homogenate waskept at room temperature for 1-1.5 h. Thirtymicroliters of this solution were used for Lowryprotein determination. After 45 min the absorp-tion was measured with a Beckman spec-trophotometer at a wavelength of 750 nmagainst a blank of NaOH solution.

2.2.8. Location of nurse beesand foragers

On the morning of day 2 of the experimentprior to the onset of foraging, the location of themarked nurse bees and pollen foragers wasnoted. Over the course of several minutes, thehive was scanned systematically, and the loca-tion of each bee was indicated with a symbol onthe glass wall of the observation hive. Thebrood areas were also outlined. This tracingwas then transferred onto another pane of glass,and photographed against a white background.A figure was prepared from the photograph byscanning it into a Macintosh computer.

3. RESULTS

3.1. Behavioral response of

pollen foragers to pollensupplementation

Figure 1 shows the total number ofpollen foragers marked on day 1 and day2 in the pollen-supplemented and thepollen-deprived colonies in each experi-ment. The graphs show that pollen sup-plementation on day 1 results in a decreaseon day 2 in the number of foragers forag-ing for pollen. The graphs along withtable I indicate that in the presence of sup-plemental pollen, a significant proportionof the pollen foragers that foraged on day1 do not forage on day 2 (in each experi-ment, P < 0.01 by a G-test of indepen-dence). Note that the pollen foragers onday 2 consist of two subgroups: 1) thosethat foraged for pollen on day 1 and whichcontinued to forage for pollen on day 2,and 2) new pollen foragers that did not

forage for pollen on day I. In the pollen-deprived colonies, the proportion of pollenforagers on day 2 compared to day 1 is

nearly 1 (0.93, 0.87 and 1.10 in experi-ments 1, 2 and 3, respectively). In con-trast, in the colonies given supplementalpollen, many fewer bees foraged for pollenon day 2 so that the proportion of pollenforagers on day 2 compared to day 1 wasconsiderably less than 1 (0.62, 0.55 and0.77 in experiments 1, 2 and 3, respec-tively). In the pollen-supplementedcolonies, two factors account for thedecrease in the number of pollen foragers.First, there are fewer pollen foragers thatcontinued to forage for pollen on day 2after having foraged for pollen on day 1.

Second, there is less recruitment of newpollen foragers. The second factor playsthe greater role in the decrease in the num-ber of pollen foragers marked on day 2.

3.2. The effect of pollensupplementation on trophallactictransfer within the colony

We wondered whether the behavioraldifferences between the pollen foragers

in the pollen-supplemented and the pollen-deprived colonies were correlated withdifferences in trophallactic transfer of pro-tein between pollen processors and pollenforagers. In all three experiments, thecolony that was treated with pollen had agreater amount of 14C-phenylalanine trans-ferred to the pollen foragers marked onday 1 than the colony which was pollendeprived (figure 2). In the three experi-ments, the amount of label transferred

(measured in DPMs) in the colony givensupplemental pollen was 2.25, 6.27 and1.58 times, respectively, the amount trans-ferred in the pollen deprived colony.

3.3. Contact between nursesand pollen foragers

Of course, in order for trophallacticinteractions to occur between pollen for-agers and pollen processors, the twogroups of bees need to have contact withone another. Therefore, we documentedthe locations of the marked pollen for-agers and the marked pollen processorson the morning of day 2. Figure 3 showsthe results for experiment 3; it provides a

momentary ’snapshot’ of the potentialinteractions between these two groups ofbees. Since the results from experiment 1were similar, only a representative figurefrom experiment 3 is presented. (Datawere not available for experiment 2.) Asseen in the figure, most of the markedpollen processors were found on the upperand middle frames which contained the

brood; although many pollen foragers wereon the lower broodless frame, many arealso on the middle frame interspersedamong the pollen processors. In addition,in the pollen-deprived colony, a greaterproportion (63 %) of the pollen foragersare on the two upper brood frames thanin the pollen supplemented colony (37 %). This is a significant difference in propor-tions (P < 0.01) based upon the G-test ofindependence. The effect was equally pro-nounced in experiment 1, where 25 % ofthe pollen foragers were seen on the twoupper brood frames in the pollen deprivedcolony versus 0 % in the pollen supple-

mented colony. Clearly, the pollen for-agers and pollen processors are in prox-imity with one another, permitting trophal-lactic interactions to occur.

3.4. The effect of pollensupplementation on midgutweight and midgut contentof foraging bees

We suspected that the behavioral dif-ferences between the pollen foragers fromthe pollen-supplemented and the pollen-deprived colonies might be correlated withphysiological differences between the beesin the two treatment groups. Therefore, asample of bees marked as pollen foragerson day 1 were collected on the morningof day 2 of the experiment as they werereturning to the observation hive. Beesreturning with pollen (pollen foragers) aswell as bees returning without pollen (nec-tar foragers) were collected. There were

10-12 bees in each group. As seen intable II, in three out of four cases, themean midgut protein content was signifi-cantly greater (P < 0.05 by the Mann-Whitney U-test) in foraging bees from thepollen-supplemented colonies. (Althoughthe fourth case showed the same effect,the variance was too large for a signifi-cant effect.) These differences in midgutprotein content can not be attributed topollen consumption, however, since themidguts of bees in both groups containedlittle pollen , and the midgut weights werenot significantly different for any pair(table II).

4. DISCUSSION

For a colony of honey bees to preciselyregulate its pollen foraging, individual for-agers must acquire information about thecolony’s nutritional needs. A potentialsource of information for the foragers isthe nurse bees which act as the pollen pro-cessors in the colony. As the primary con-sumers and dispensers of both pollen and

proteinaceous jelly derived from pollen,these bees automatically integrate infor-mation about the colony’s supply anddemand for pollen in the course of theirdaily activities: information about pollensupply is gathered as they search for, andconsume pollen; information about pollendemand is available as they provide pollenand jelly to brood and adult bees whichneed proteinaceous nourishment.

The present study supports the hypoth-esis that the pollen processors (nurse bees)in the colony provide pollen foragers withreliable information concerning thecolony’s need for pollen which they useto modulate their level of pollen foragingactivity. The evidence is as follows: thereis contact between these two groups ofbees allowing opportunity for trophallac-tic interactions to occur (figure 3). In addi-tion, compared to conditions of pollenexcess, under conditions of pollen depri-vation, a greater proportion of foragersare found in the brood nest where they aremore likely to encounter nurse bees (fig-ure 3). Under conditions of pollen sup-

plementation, pollen processors distributemore labeled protein to foragers thanpollen processors in colonies deprived ofpollen (figure 2). The greater amount oflabel transferred to the marked pollen for-agers in the pollen supplemented coloniessuggests that pollen processors either havemore protein to dispense or that they aremore willing to dispense that protein, orboth.

We suggest that these differences introphallactic transfer of jelly provide cuesto the pollen foragers which are used tomodulate their foraging activity, thusaccounting for the observed differencesin the foraging behavior of the pollen-sup-plemented versus pollen-deprived foragers(figure 1). Foragers do not need to directlyassess the pollen stores (Camazine, 1993),nor do they obtain information about thecolony’s nutritional status by consumingpollen themselves. Midgut weights didnot increase significantly in the bees fromthe pollen-supplemented colonies (table II)as would be expected if they ingested morepollen. Furthermore, pollen foragers havebeen shown not to eat significant amountsof pollen (Crailsheim et al., 1992), corre-lated with a decrease in the proteolyticactivity of their gut enzymes (Moritz andCrailsheim, 1987). Nonetheless, pollenforagers from colonies given supplemen-tal pollen do show physiological changesindicating increased protein intake: theprotein content of their gut is greater(table II) and they have higher body lev-els of radio-labeled protein than pollenforagers from colonies deprived of pollen(figure 2). Presumably the additional pro-tein was obtained through trophallacticinteractions as jelly fed to the foragers bythe pollen processors.

Why have we hypothesized this indi-rect mechanism of information acquisi-tion rather than an apparently simpler, andequally reliable mechanism in which thepollen foragers themselves assess thecolony’s need for pollen? One argument is

that direct assessment would presumablyrequire extensive and ongoing surveys byeach forager of both colony’s supply(pollen stores) and demand (amount ofbrood and their state of nourishment).Such a feat of information collectionwould undoubtedly be time consuming,if not impossible, and would certainlydetract from the foragers primary task ofgathering pollen.

A second argument against directassessment comes from previous experi-ments (Camazine, 1993) which showedthat pollen foragers do not require directcontact with the colony’s pollen stores toassess colony pollen needs. Pollen for-agers separated from the colony’s pollenstores by a double screen (which preventstrophallactic interactions between foragingbees and house bees) appeared to lackinformation about the colony’s pollen sta-tus. In contrast, foragers separated by asingle screen (which prevented access topollen stores, but allowed trophallacticinteractions across the screen) obtainedaccurate information about the colony’spollen status and foraged appropriately.These experiments suggested that a non-foraging class of bees was able to indi-rectly provide the foragers with reliableinformation concerning the colony needfor pollen.

Such a system of indirect information

acquisition may be both effective and effi-cient. All the pollen forager needs to doin order to accurately assess the colony’sneed for pollen is to sense her own hungerfor protein. The system is similar to thatfor the regulation of nectar foraging (See-ley, 1989), where the nectar foragers mod-ulate their foraging behavior based uponsimple cues obtained from the nectarreceiving class of bees. Here, too, the for-agers do not, themselves, directly assessthe colony’s honey stores, but do so indi-rectly through information provided byother bees which automatically obtain that

information in the course of their nectar

storing activities.

Although the results presented here sup-port the hypothesis that trophallactic inter-actions (transfer of proteinaceous jellyfrom pollen processors to pollen foragers)provide information used by the foragersto modulate their pollen foraging, severalimportant questions remain. Foremost isan explanation of precisely how trophal-lactic interactions provide information tothe pollen foragers. We can suggest sev-eral alternatives: 1) the ’amount’ of proteinreceived by a pollen forager from a pollenprocessor may affect the forager’s likeli-hood of foraging, 2) The ’ease’ with whicha pollen foragers receives protein from apollen processor may affect the foragerslikelihood of foraging. Ease might be eval-uated by the pollen forager as a function ofthe ’number’ of interactions required toreceive a specific amount of protein froma pollen processor, the ’time’ required toreceive a specific amount of protein, orthe number (’proportion’) of trophallac-tic encounters with pollen processors inwhich the pollen forager is able to receivejelly.

With each of these proposed mecha-nisms, under conditions of ample pollenreserves, a pollen forager would presum-ably find it easy to locate pollen processorsable and willing to transfer large amountsof proteinaceous jelly. Despite the plau-sibility of this hypothesis, it still remainsto be demonstrated that trophallactic inter-actions actually modulate the behavior ofthe forager. In addition, if this is the mech-anism, there remains the question ofwhether it is the physiological effects ofthe transfer of jelly that is important, orwhether associated behavioral interactionsthat take place during trophallaxis are cru-cial in affected subsequent foraging activ-ity? It might be possible to answer thesequestions if one were able to artificiallyfeed proteinaceous jelly or jelly compo-nents to cohorts of foragers under con-

trolled conditions and observe their sub-

sequent foraging behavior. Such an exper-iment would provide much-needed directevidence for the role of protein trophal-laxis in the regulation of pollen foraging.

ACKNOWLEDGMENTS

We would like to thank the many individu-als who contributed technical assistance on this

project, including Barbara Berger, RosannaGiordano, Tugrul Giray, Dan Guyot, JackKuehn, Maximillian Marek, Ute Panzenboeck,Karl Pfeiffer, Leo Schneider, Birgit Schoeff-mann, Anton Stabentheiner, Susan Trainor,Josef Ulz, Maria Wedenig and Ginger With-ers. Special thanks to Tom Seeley, Kirk Viss-cher, Stefan Fuchs and an anonymous reviewerfor their helpful comments and suggestions.Supported, in part, by a grant from the Aus-trian Science Foundation (to KC), grants fromthe USDA and NIMH (to GER), a grant fromthe Fogarty Foundation of NIH (to GER andKC), and a National Science Foundation Doc-toral Dissertation Improvement Grant (to SC).

Résumé - Trophallaxie des protéines etrégulation de la récolte du pollen parles abeilles (Apis mellifera L.). Une colo-nie d’abeilles régule de façon précise sarécolte de pollen. Pour ce faire, les buti-neuses de pollen doivent acquérir chacunedes informations concernant les besoinsnutritionnels de la colonie. Une étude anté-rieure (Camazine, 1993) a montré qu’iln’était pas nécessaire que les butineusesde pollen aient un contact direct avec lesprovisions de pollen pour connaître lesbesoins de la colonie. Au contraire, l’infor-mation semble être obtenue indirectementà partir des autres individus de la colonie.Les nourrices constituent une source

potentielle d’information puisqu’elles sontles premières transformatrices du pollenen le consommant, en le convertissant

rapidement en gelée (sécrétion protéiniquede la glande hypopharyngienne) et en ledistribuant aux adultes et aux larves danstoute la colonie. Nous avons étudié le

transfert des protéines (trophallaxie) parles nourrices, transformatrices de pollen,comme source d’informations pour lesbutineuses. Nous avons testé l’hypothèseselon laquelle des variations dans le fluxde gelée protéinique en direction des buti-neuses de pollen serviraient d’indicationsutilisées par ces dernières pour estimer lesbesoins protéiniques de la colonie ; cesvariations leur fourniraient des informa-tions qui leur permettraient de décider sielles doivent ou non poursuivre la récoltede pollen. L’expérimentation a porté surdeux colonies en parallèle ; toutes deuxont été privées de pollen durant quelquesjours, puis l’une a reçu du pollen tandisque l’autre en est restée privée.

Les résultats sont les suivants : les buti-neuses de pollen perçoivent très rapide-ment - en quelques heures - les change-ments dans les provisions de pollen de lacolonie. Dans la colonie supplémentée, denombreuses butineuses de pollen réagis-sent en ne poursuivant pas leur récolte.Soit elles se mettent à récolter du nectar,soit elles cessent toute activité de butinage(figure 1 et tableau I). Par ailleurs, selonqu’elles viennent des colonies supplé-mentées ou des colonies privées de pol-len, les nourrices n’ont pas le même com-portement vis-à-vis des butineuses de

pollen. Elles réagissent à la supplémenta-tion en pollen en donnant aux butineusesune grande partie de la gelée nouvelle-ment synthétisée. Les nourrices des colo-nies supplémentées, auxquelles on ainjecté de la phénylalanine marquée auC14, distribuent plus de protéines mar-quées aux butineuses que celles des colo-nies non supplémentées (figure 2). Dansles trois expériences, la quantité de nour-riture marquée transferrée aux butineuses(mesurée en désintégrations par minute)dans la colonie supplémentée était res-pectivement de 2,25, 6,27 et 1,50 fois celletransferrée dans la colonie non supplé-mentée. Comme le montre le tableau III,dans trois cas sur quatre, la teneur

moyenne en protéines de l’intestion moyenétait significativement plus élevée

(p < 0,005 test-U de Mann-Whitney) chezles butineuses des colonies supplémen-tées. Ces différences ne peuvent être attri-buées à la consommation de pollen par lesbutineuses de pollen elles-même puisqueles intestins moyens des abeilles des deux

groupes renfermaient peu de pollen et quele poids des intestins moyens ne différaitsignificativement pour aucune des paires(tableau II). Ces résultats suggèrent aucontraire que les butineuses de pollenobtiennent des protéines en sollicitant dela nourriture auprès des nourrices. Ceciconfirme l’hypothèse selon laquelle lesinteractions trophallactiques entre nour-rices et butineuses de pollen servent d’indi-cations informant ces dernières des besoinsen pollen de la colonie. &copy; Inra/DIB/

AGIB/Elsevier, Paris

Apis mellifera / trophallaxie / protéine /nourrice / récolte pollen

Zusammenfassung - Proteintrophalla-xis und die Regulation des Pollensam-melns bei Honigbienen (Apis melliferaL.). Honigbienenvölker können den Ein-trag von Pollen sehr genau regulieren.Hierzu müssen die einzelnen Pollen-sammlerinnen notwendigerweise Infor-mation über die Ernährungslage des Bie-nenvolkes erhalten. In früheren

Untersuchungen (Camazine, 1993) konntebereits gezeigt werden, da&szlig; die Pollen-sammlerinnen zur Abschätzung des Pol-lenbedarfs keinen direkten Kontakt zu denPollenvorräten haben müssen. Stattdessenscheint die Information indirekt von ande-ren Arbeiterinnen im Volk zu stammen.Eine der möglichen Informationsquellenstellen hierbei die Ammenbienen als die

primären Pollenverarbeiterinnen im Bie-nenvolk dar. Diese verzehren den Pollenund wandeln ihn sehr rasch in das protein-haltige Sekret der Hypopharynxdrüsenum, das sie sowohl an die Larven als auch

an die ausgewachsenen Tiere im ganzenVolk weitergeben. In der vorliegendenStudie wird die mögliche Rolle des Trans-fers von Protein (Trophallaxis) durch diesePollenverarbeiterinnen (Ammenbienen)als Informationsquelle für die Sammel-bienen untersucht. Wir prüften die Hypo-these, da&szlig; ein unterschiedlich starker Flu&szlig;von proteinhaltigem Drüsensekret von denAmmenbienen zu den Sammelbienen vondiesen als Hinweis auf den Proteinbedarfdes Volkes genutzt wird, auf Grund dessensie entscheiden, ob sie weiterhin Pollensammeln. Wir berichten hier über folgendeErgebnisse: Pollensammlerinnen nehmeneine Veränderung der Pollenvorräte inner-halb weniger Stunden, also sehr raschwahr. In einem Volk, dem nach Pollen-deprivation zusätzlicher Pollen gegebenwird, unterbrechen viele der Pollen-sammlerinnen den Polleneintrag . Siegehen dann entweder zum Nektarsammelnüber, oder sie stellen die Sammelflügeganz ein (Figure 1 und Tabelle I). In Ver-bindung mit dieser Änderung des Sam-melverhaltens zeigen die Pollenverarbei-terinnen in den mit zusätzlichem Pollen

versorgten Völkern ein anderes Verhal-ten gegenüber den Pollensammlerinnenals in den pollendeprivierten Völkern. Siereagieren auf die Zufügung von Pollen mitder Übertragung einer grö&szlig;eren Mengeihres frisch synthetisierten Futtersaftei-wei&szlig;es. Mit 14C-phenylalanin injizierteProteinverarbeiterinnen in den mitzusätzlichem Pollen versorgten Völkernverteilen mehr markiertes Eiwei&szlig; anSammlerinnen als entsprechende Pollen-verarbeiterinnen in Völkern ohne zusätz-lichen Pollen (Figure 2). In den drei Expe-rimenten war die Menge des übertragenenMarkers (gemessen in DPMs) in dem Volkmit zusätzlichem Pollen 2.25, 6.21 bezie-hungsweise 1.58 mal so hoch wie in denmit Pollen unterversorgten Völkern.Tabelle II zeigt, da&szlig; in drei von vier Fällender mittlere Proteingehalt im Mitteldarmbei den Sammlerinnen in mit zusätzlichemPollen versehenen Völkern signifikant

erhöht war (P < 0.05, Mann-Whitney Utest). Diese Unterschiede können nicht aufden Verzehr von Pollen durch die Samm-lerinnen selbst zurückgeführt werden, da inbeiden Gruppen der Pollengehalt im Mit-teldarm nur sehr gering war und dieGewichte des Mitteldarms bei keinem derPaare unterschiedlich war (Tabelle II).Das Ergebnis weist darauf hin, da&szlig; diePollensammlerinnen sich von den Pollen-verarbeiterinnen mit Protein versorgen las-sen. Diese Befunde unterstützen daher die

Hypothese, da&szlig; trophallaktische Interak-tionen mit den Pollenverarbeiterinnen denPollensammlerinnen als Hinweis zur Ein-

schätzung der Pollenversorgung des Bie-nenvolkes dienen. &copy; Inra/DIB/AGIB/

Elsevier, Paris

Apis mellifera / Pollensammlerinnen /Proteintrophallaxis / Ammenbienen

REFERENCES

Al-Tikrity WS, Benton AW, Hillman RC, ClarkeWW Jr. (1972) The relationship between theamount of unsealed brood in honeybee coloniesand their pollen collection. J Apic Res 11, 9-12

Barker RJ (1971) The influence of food inside thehive on pollen collection by a honeybee colony.JApic Res 10, 23-26

Cale GH Jr (1968) Pollen gathering relationship tohoney collection and egg-laying in honey bees.Am Bee J 108, 8-9

Camazine S (1993) The regulation of pollen foragingby honey bees: how pollen foragers assess thecolony’s need for pollen. Behav Ecol Sociobiol32, 265-272

Crailsheim K (1986) Dependence of proteinmetabolism on age and season in the honeybee(Apis mellifica carnica Pollm). J Insect Physiol32, 629-634

Crailsheim K (1990a) The protein balance of thehoney bee worker. Apidologie 21, 417-429

Crailsheim K (1990b) Protein synthesis in the hon-eybee (Apis mellifera L.) and trophallactic dis-tribution of jelly among imagos in laboratoryexperiments. Zool Jb Physiology 94, 303-312

Crailsheim K (1991) Interadult feeding of jelly inhoneybee (Apis mellifera L.) colonies. J CompPhysiol B 161, 55-60

Crailsheim K, Stolberg E (1989) Influence of diet,age and colony condition upon intestinal prote-

olytic activity and size of the hypopharyngealglands in the honeybee (Apis mellifera L.). JInsect Physiol 35, 595-602

Crailsheim K, Schneider LHW, Hrassnigg N,Buehlmann G, Brosch U, Schöffmann B, Gmein-bauer R (1992) Pollen consumption and utiliza-tion in worker honey bees (Apis mellifera car-nica): dependence on individual age and function.J Insect Physiol 38, 409-419

Fewell JH , Winston ML (1992) Colony state andregulation of pollen foraging in the honey bee,Apis mellifera L. Behav Ecol Sociobiol 30,387-393

Free JB (1967) Factors determining the collectionof pollen by honeybee foragers. Anim Behav 15,134-44

Free JB, Williams IH (1971) The effect of givingpollen and pollen supplements to honeybeecolonies on the amount of food collected. J ApicRes 10, 87-90

Heinrich B (1981a) Energetics of honeybee swarmthermoregulation. Science 212, 565-566

Heinrich B (1981b) The mechanisms and energet-ics of honeybee swarm temperature regulation. JExp Biol 91, 25-55

Heinrich B (1981c) The regulation of temperaturein the honeybee swarm. Sci Am 244, 146-160

Lindauer M (1952) Ein Beitrag zur Frage der Arbeits-teilung im Bienenstaat. Z Vergl Physiol 34,299-345

Lindauer M (1954) Temperaturregulierung undWasserhaushalt im Bienenstaat. Z Vergl Phys-iol 36, 391-432

Lowry OH, Rosebrough NJ, Farr AL, Randall RJ(1951) Protein measurement with the Folin phe-nol reagent. J Biol Chem 193, 165-275

Lüscher M (1961) Air-conditioned termite nests. SciAm 205, 138-145

Moeller FE (1972) Honey bee collection of cornpollen reduced by pollen feeding in hive. Am BeeJ 112, 210-212

Moritz B, Crailsheim K (1987) Physiology of proteindigestion in the midgut of the honeybee (Apismellifera L.). J Insect Physiol 12, 923-931

Seeley TD (1989) Social foraging in honey bees:how nectar foragers assess their colony’s nutri-tional status. Behav Ecol Sociobiol 24, 181-199

Seeley T, Towne W (1992) Tactics of dance choicein honey bees: do foragers compare dances?Behav Ecol Sociobiol 30, 59-69

Seeley TD, Camazine S , Sneyd J (1991) Collectivedecision-making in honey bees: how colonieschoose among nectar sources. Behov Ecol Socio-

biol 28, 277-290

Sorensen AA, Busch TM, Vinson SB (1985) Controlof food influx by temporal subcastes in the fireant, Solenopsis invicta. Behav Ecol Sociobiol 17,191-198

Tschinkel WR (1988) Social control of egg-layingrate in queens of the fire ant, Solenopsis invicta.Physiol Ent 13, 327-350

Van Laere O , Martens N (1971) Influence d’unediminution artificielle de la provision de pro-téines sur l’activitié de collete de la colonied’abeilles. Apidologie 2, 197-204