HAL Id: hal-00891138https://hal.archives-ouvertes.fr/hal-00891138

Submitted on 1 Jan 1994

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.

Reproductive cycling and hierarchical competition inCape honeybees, Apis mellifera capensis Esch

Hr Hepburn

To cite this version:Hr Hepburn. Reproductive cycling and hierarchical competition in Cape honeybees, Apis melliferacapensis Esch. Apidologie, Springer Verlag, 1994, 25 (1), pp.38-48. <hal-00891138>

Original article

Reproductive cycling and hierarchical competitionin Cape honeybees, Apis mellifera capensis Esch

HR Hepburn

Department of Zoology and Entomology, Rhodes University, PO Box 94,6140 Grahamstown, South Africa

(Received 16 February 1993; accepted 24 June 1993)

Summary &mdash; On queen loss, Cape honeybees may requeen from the brood of a queen, an egg-laying worker or it may remain as a laying worker colony. These 3 possibilities are reached through8 different social developmental pathways. The fate of any individual queenless colony varies withbrood status and the rates and kinds of worker differentiation that occur on queen loss. Worker diffe-rentiation is widespread and includes ovarial and pheromonal development; but only few individualsactually reach "surrogate queen" status. The complex pathways that occur in queenless Cape colo-nies seem to be determined both by social context and hierarchical competition among workers.

Apis mellifera capensis / reproduction / surrogate queen / egg-laying worker / hierarchy /pheromone

INTRODUCTION

Hierarchical competition for reproductivedominance is continually suppressed bypheromones of the honeybee queen (cfFree, 1987). On her loss, there is a strug-gle that includes fighting, development ofthe ovaries and more queenlike phero-mones in the workers (Velthuis et al,1990). These changes are further modu-lated by brood and queen cells (Jay, 1968,

1970). In queenless Cape honeybee colo-nies hierarchical competition is a skein ofcontinually changing and reciprocally inter-active variables (Hepburn et al, 1988; Vel-thuis et al, 1990). In fact, queenless Capebees may follow at least 8 different path-ways of social development. These wereexamined with respect to brood and the

pheromonal and ovarial changes that oc-cur at critical periods in the social develop-mental pathways on queen loss.

MATERIALS AND METHODS

Over a 5-yr period 95 colonies of Apis melliferacapensis averaging about 10 000 bees/colonywere collected at Grahamstown, hived, and al-lowed to settle in an apiary. All were de-

queened, some were kept broodright (n = 39)while others were also debrooded (n = 56) andtheir developmental pathways recorded. Finally,17 colonies were dequeened, 5 of which hadtheir queens’ brood and 12 were debrooded.Five hundred worker bees were collected fromeach of these 17 colonies on dequeening and160 bees/colony for each subsequent week, theduration of sampling depending on colony fate.Each bee’s head was placed in dichlorome-thane and its matching and coded abdomenwas frozen.

Workers were scored for ovarial develop-ment using a slight modification of Velthuis

(1970): 0 = ovaries undeveloped, no compart-mentation; 1 = compartmentation visible, but nooocyte development; 2 = ovarioles swollen, oo-cytes spherical; 3 = oocytes bean-shaped; 4 =oocytes sausage-shaped, ripe. The head ex-

tracts were derivatized with BSTFA, concentrat-ed with nitrogen and analyzed with a HewlettPackard 5890 Series II gas chromatograph fit-ted with a fused silica column. Six compoundswere measured against authentic samples andinternal standards: 8HOA = 8-hydroxyoctanoicacid; 9ODA = 9-oxo-2-decenoic acid; 9HDAA =

9-hydroxydecenoic acid; 9HDA = 9-hydroxy-2-decenoic acid; 10HDAA = 10-hydroxydecenoicacid and 10HDA = 10-hydroxy-2-decenoic acid.At least 50 bees were analyzed for each weekof each pathway shown in figure 1. Pheromonalclasses are defined as the relative percentageof 9ODA to the other 5 assayed components ofthe mandibular gland extract: class 1 = 0-33.3%9ODA, 2 = 33.4-66.6% 9ODA and 3 = 66.7-100% 9ODA.

Working definitions of workers are assignedas follows: normal workers = bees in stages 0-1of ovarial development with a worker-like 9ODAcontent less than 1/3 of the bouquet compoundsmeasured; laying workers = bees in stage 4,9ODA content less than 1/3; false queens =

bees in stages 0-3, 9ODA exceeds 2/3; surro-gate queens = bees of stage 4, 9ODA exceeds2/3; intermediates = all other permutations.(Note: the categories of workers are defined

solely on ovarial and pheromonal content. Their

precise behaviours, including whether every beewith ripe eggs actually lays eggs, are not yetknown.) Frequency values (tables I and II) arepercentages of a kind of bee for the total samplesize of its group per specified week. These datawere analyzed with the Mann-Whitney U-testsince all data were at least at the ordinal level ofmeasurement. The test statistic is z, not the usu-al U, because sample sizes were always great-er than 20. Significance is defined as p < 0.05.Statistical comparisons of the developmentalpathway events were made using the Kruskal-Wallis test applied to values from an analysis ofvariance.

RESULTS

Reproductive cycles

Four distinct pathways were observed in

queenless broodright colonies of the Capebees (P1-4, fig 1). The colonies may con-struct queen cells from available queenbrood and simply rear a new queen("classical requeening", pathway 2), or be-gin rearing a queen, but abort the wholeprocess only to repeat it with the eggs of aworker (pathway 1). Thirdly, a colony maynot construct queen cells but persist forseveral months only with laying workersand/or surrogate queens (pathway 3), butprobably eventually dwindle. Lastly, the

colony may abscond (pathway 4), desert-ing the brood, in which case it enters thebroodless pathways (fig 1).

Broodless colonies of Cape bees be-have quite differently. If workers become

egg-layers then the colonies may persist along time (pathway 5) just as old coloniesdo along pathway 3. Colonies whose work-ers produce brood may begin to constructqueen cells and workers may producesome offspring but the colony soon dwin-dles (pathway 7) or absconds. Finallysome colonies never produce worker

brood and either abscond or dwindle (path-way 8).

On dequeening (week 0) the broodrightand the broodless colonies did not differ

significantly in median ovarial developmentnor were there any laying workers, falsequeens or surrogate queens. A medianovarial class of 0 persisted throughout theexperiment and indicates that changes oc-curred among relatively few bees. A weeklater all colonies (fig 1, table I) containedfalse queens, most had laying workers (ex-cept pathway-4 bees) and surrogatequeens (except those of pathway 1, not

considering the absconded and dwindledcolonies). These developments occurreddespite the presence of queen cells in

pathways 1 and 2 and queen brood in

pathways 1-3 and worker brood in 6-7.

By the end of week 2, pathway 1 colo-nies (requeening from laying-workerbroods) destroyed previously constructedqueen cells. All old queen brood had be-come capped and all colonies containednew young brood derived from workers (fig1, table I). At week 3 all colonies contained

mixed-age brood derived from workers andthere were capped queen cells in path-ways 1, 2 and 6. The percentages of layingworkers, false queens and surrogatequeens were significantly elevated com-

pared to week 2. By week 4, bees of path-ways 1, 2 and 6 had either virgin or matedqueens. The other colonies had eitherdwindled or absconded. Now there were

significant reductions in the percentages oflaying workers, false queens and surrogatequeens (table I). By the end of week 5,pathway 1, 2 and 6 colonies had returnedto ’normal status’ and were headed by mat-ed, laying queens and the reproductiveworkers had disappeared (fig 2).

In these developmental sequences only15 colonies (15.8%) followed that of"classical requeening" (fig 1, pathway 2).Other queen production was based on

worker eggs (fig 1, pathways 1 and 6,20%). As a general trend the relative per-centages of laying workers, false queensand surrogate queens all increased fromweek 0 to week 3 and then all went intodecline with the appearance of newly mat-ed queens. However, in pathway 2 (classi-cal requeening) the cumulative frequencyof egg-laying workers was significantly lessthan in colonies that requeened from work-er brood (table I). In pathway 6 (requeen-ing from worker’s progeny) up to 15% ofthe workers were potentially reproductive.

Hierarchical competition

Reproductive differentiation among queen-less worker bees with changing colonycontext is given in table II. On dequeeningall colonies were queenright and brood-right. Only a few bees showed slight ovari-al or pheromonal development. Half of thecolonies retained their mother’s brood

(group 1 = pathways 1 and 2), the otherhalf were debrooded (group 2 = pathways4, 6 and 7). Samples collected 7 d later

(week 1) significantly differed from those ofweek 0 (Z = 20.9 for OD, Z = 5.5. for

9ODA) but not from each other as groups.Within a week both groups had developedlaying workers, false queens and surrogatequeens as well as intermediates (table II).

As the colonies entered the secondweek of queenlessness the context of

group 1 (pathways 1 and 2) had changedto contain brood, queen cells and differen-tiated workers; that of group 2 (pathways4, 6 and 7) included the brood of layingworkers and/or surrogate queens as wellas differentiated workers. All colonies werenow broodright but group 1 colonies had

more brood than group 2. At week 2, group2 colonies also had developing queensand there was a significant increase in lay-ing workers compared to week 1 (Z = 3.8for OD). There was no significant changein worker differentiation in group 1 be-

tween weeks 1 and 2 (table II).

During the third week of queenlessness,contextual differences between groups 1

and 2 dissipated. The original queen’sbrood of group 1 had matured and

eclosed; its remaining brood was (likegroup 2) that produced by laying workersand/or surrogate queens. By the next sam-ple time (week 3 = pathways 1, 2, 4, 6 and7) all colonies had converged on the samedevelopmental pathway and contained ei-ther very ripe queen cells or just eclosedvirgin queens. Group 1 reared queensfrom their original queen’s progeny, group2 from the progeny of laying workers orsurrogate queens. Because the social con-text for both groups was now virtually iden-tical they were combined. At week 3 (tableII) worker differentiation had virtually dou-bled. There were many intermediates and

significantly more laying workers (Z = 12.1)and significantly more pheromonal change(Z = 0.45).

In the ensuing week, the developing orjust emerged young queens had mated

just prior to sampling at week 4. Now therewas a significant reduction in ovarial devel-opment (Z = -6.7) and in pheromonal de-velopment (Z = -10.7) compared to week3 (table II). These workers were largely in-termediates, the process of reproductivedifferentiation in workers now being re-

versed. By the 5th week all of the colonieswere classically queenright and headed bymated, laying queens. At the last samplinginterval (week 5) worker differentiation hadsignificantly regressed in both ovarial (Z =-2.8) and pheromonal (Z = -2.9) terms.

DISCUSSION

The reproductive options open to Capebees on queen loss are classical requeen-

ing, requeening from worker brood or re-maining as a laying-worker colony. About60% of the broodright colonies requeenedfrom either queen’s or worker’s brood leav-ing 40% to proceed as laying worker colo-nies (fig 1). Only 20% of the colonies suc-cessfully requeened from laying workerbrood. Nearly 40% of the queenless colo-nies absconded (about 20% of the brood-right and 50% of broodless colonies). Thefates of these colonies are not known.Nevertheless, thelytoky in Cape workerbees is a regular reproductive pathway forthis race. The frequencies of this and otherpathways followed is related to brood con-text and possibly to seasonal influences.Almost all of the data (fig 1) is derived fromfall and winter, when queenless Capeswarms are commonplace and colonies

abscond leaving all stages of brood behind(Hepburn et al, 1988).

Turning to the social milieu of the

queenless nest, broodrightness and brood-lessness in the Cape honeybee are radi-cally different from the same contexts in

other races of honeybees (Ruttner andHesse, 1979). In queenless Cape honey-bees, uncapped brood stimulates queen-cell construction and partially inhibits ovari-al development in workers (Hepburn et al,1988). If queen rearing proceeds, further

queen-cell construction is inhibited as is re-productive differentiation among workers

(Whiffler and Hepburn, 1991). As more ofthe brood of the former queen is capped,reproductive differentiation of the workers

becomes derepressed which in turn inhib-

its queen-cell construction and may eveninclude the destruction of existing queencells (Hepburn et al, 1988).

Thus there is competition involving thedeveloping queens and the differentiatingworkers (fig 2). If a queen emerges and

mates, she inhibits queen-cell construc-

tion, worker differentiation and causes dif-ferentiated workers to regress, all in onefell swoop (Whiffler and Hepburn, 1991).If, however, workers have differentiated farenough toward surrogate queenship priorto emergence of a queen, queen rearingmay be abandoned and, if broodless, thecolony may enter pathways of the brood-less context (figs 1 and 2, Hepburn et al,1988). In queenless colonies the absenceof brood and developing queen cells dere-presses ovarial development (Allsopp,1988), leads to changes in the pheromonalbouquet of the mandibular gland (Hemm-ling et al, 1979), and the development ofhierarchical dominance (Hillesheim et al,1989) among workers. These race-specifictraits (Ruttner and Hesse, 1979) have dif-ferent heritability indices (Moritz and Hille-sheim, 1985) and operate under differentconstraints.

The data of table II substantiates this in-

terpretation but show that the rates ofovarial and pheromonal development andsubsequent regression are not the same.On dequeening (week 0) there are no ripeeggs or queenlike pheromones among theworkers. In group 1, 11% of the bees have

ripe eggs after a week and even up to theemergence of the queen and her mating,11% of the bees have ripe eggs. The

pheromonal picture is quite different: thequeenlike bouquet rises quickly in weeks 1

and 2 but then regresses very quickly withthe emergence of the queen. In group 2,more bees have ripe eggs than the beesof group 1 and the queenlike pheromonesregress far more slowly (table II). In theabsence of brood ovarial development iseither stimulated or derepressed yetqueenlike bouquets are more prevalent inthe presence of brood (group 1) than in itsabsence (group 2).

Ovarial development is inversely relatedto age and increases in frequency with ex-tended absence of repressive factors

(Hepburn et al, 1988, 1991). Both the bou-quet (Hemmling et al, 1979) and theamount of 9ODA (Crewe 1988) in the man-dibular gland varies with age, resulting in awide pheromonal spectrum of bees fromworkerlike to queenlike (Hepburn, 1992).Ovarial development and becoming phe-romonally queenlike covary (Hepburn,1992), and, with differentiation, a domi-

nance hierarchy also develops amongthem (Hillesheim and Moritz, 1987) in

which dominants are attacked by other

bees (Velthuis et al, 1990). Broodlessness,coupled to the above leads to worker beesof varying status, the frequencies of whichvary with changing context (tables I and II;figs 1 and 2). If colonies are queenless forlong enough (a couple of weeks) they de-velop an entirely new range of stimulatoryand/or inhibitory properties (Hepburn et al,1988, 1991) and these are represented bynormal workers, laying workers, false

queens, surrogate queens and intermedi-ates (table II).

Worker bees that remain "normal" at-

tack differentiating bees (Velthuis and vander Kerk, 1988), are trophallactically subor-dinate and are themselves the targets ofinhibitory signals from differentiating work-ers (Hillesheim et al, 1989). Laying work-ers and surrogate queens are attacked byother workers (Velthuis and van der Kerk,1988). False queens inhibit queen-cellconstruction (Whiffler and Hepburn, 1991).worker ovarial development (Velthuis et al,1990) and changes in the pheromonalbouquet of other workers (Hemmling et al,1979). They are also trophallacticallydominant (Hillesheim et al, 1989) andevoke retinue formation in other workers

(Velthuis, 1985). Surrogate queens havethe properties of both laying workers andfalse queens and virtually mimic real

queens (Ruttner et al, 1976; cf tables I and

II). Intermediates remain behaviourally un-assayed but may be operationally interme-diate in their effects on other bees and

how they in turn are affected by otherbees.

The sequence of possibilities modulatedby interactions among queenless Capeworker bees are schematically representedin figure 2. These appear to be the minimalphysiological and behaviourial relation-

ships required to explain the stimulatoryand inhibitory effects (fig 2) that form thebasis of hierarchical competition on queenloss in the Cape honeybee.

ACKNOWLEDGMENTS

I thank RM Crewe for advice, discussion and forthe gift of GC Standards, and HHW Velthuis andRFA Moritz for extremely useful comments onthe manuscript.

Résumé &mdash; Cycles de reproduction etcompétition hiérarchique chez l’abeilledu Cap, Apis mellifera capensis Esch. Àla perte de leur reine les abeilles du Cappeuvent soit remplacer la reine à partir ducouvain de reine ou du couvain d’ouvrières

pondeuses, soit rester orphelines (fig 1).Quatre voies de développement sont étu-diées pour des colonies orphelines maispossédant du couvain : 1) remplacementclassique de la reine par élevage d’unenouvelle reine ; 2) construction d’une cellu-le royale de sauveté à partir d’une larve dereine, qui est détruite puis remplacée parun oeuf d’ouvrière pondeuse ; 3) aucunessai d’élevage de reine, la colonie conti-nue avec des ouvrières pondeuses et/oudes reines suppléantes ; 4) les colonies

peuvent déserter et intégrer la voie de dé-veloppement sans couvain. Les coloniessans couvain peuvent continuer commepetites unités dirigées par des ouvrièrespondeuses et/ou des reines suppléantes(5) ou bien une reine peut être produite àpartir d’un oeuf d’ouvrière pondeuse (6).Les colonies sans reine ni couvain peuvent

également produire une petite descendan-ce d’ouvrières pondeuses mais dépéris-sent ou désertent rapidement (7). Finale-ment (8) aucun couvain d’ouvrière

pondeuse n’est produit et la colonie dépéri-ra ou désertera (fig 1). Les colonies qui ontdéserté peuvent suivre l’une des voies dedéveloppement sans couvain ou se joindreà d’autres essaims en vol ou qui se sontregroupés. Dans de tels cas les reines ensurnombre, les ouvrières pondeuses et lesreines suppléantes sont éliminées de lanouvelle colonie.

Le destin de toute reine varie avec la

statut du couvain et les taux et types dedifférenciation des ouvrières qui existent àla perte de la reine (fig 2). La différencia-tion des ouvrières orphelines est couranteet il existe un large spectre de différencia-tions ovarienne et phéromonale suscepti-bles de se produire (tableau I). Il comprendles abeilles qui, du point de vue phéromo-nal, sont semblables aux ouvrières et qui(1) pondent ou (2) ne pondent pas d’&oelig;ufset celles qui, du point de vue phéromonal,sont semblables aux reines ; certaines

pondent des &oelig;ufs (3), d’autres n’en pon-dent pas (4). Il existe de nombreux autresintermédiaires qui sont des permutationsde ces types. Très peu d’ouvrières se diffé-rencient en véritables «reines suppléan-tes» qui sont très proches des reines sur leplan phéromonal et qui pondent des &oelig;ufs

(tableau II). Si l’on examine les fréquencesrelatives de la différenciation ovarienne et

phéromonale chez les colonies orphelines(tableau II) en fonction des 8 voies de dé-veloppement que peuvent suivre des colo-nies orphelines (fig 1), on s’aperçoit que lanature de la différenciation des ouvrièressemble être déterminée par des interac-tions complexes, impliquant le contexte so-cial et la compétition hiérarchique entre lesouvrières orphelines (fig 2).

Apis mellifera capensis / reproduction /hiérarchie-dominance / ouvrière pon-

deuse / reine suppléante / régulationsociale / phéromone

Zusammenfassung &mdash; Zyklische Abläufein der Reproduktion und hierarchischeKonkurrenz bei der Kapbiene, Apis mel-lifera capensis Esch. Nach dem Verlustder Königin können sich Kapbienen entwe-der aus der Brut einer Königin oder einerlegenden Arbeitsbiene wiederbeweiseln,oder sie können weisellos bleiben (Abb 1).Vier Entwicklungswege für weiselloseVölker mit Brut werden dokumentiert: 1)Klassische Beweiselung durch Aufzuchteiner Nachschaffungskönigin; 2) Nach-

schaffungszellen mit Weisellarven, dieaber zerstört und mit Eiern legender Ar-beitsbienen erfolgreich zu Ende gepflegtwerden; 3) Es kommt zu keinem Versucheiner Weiselaufzucht, sondern das Volkexistiert weiter mit legenden Arbeitsbienenund/oder Ersatz- (Surrogat-)Königinnen;und 4) Die Völker können ausziehen undden Entwicklungsweg der Brutlosigkeitnehmen. Brutlose Völker können als kleineEinheiten weiter bestehen, mit legendenArbeitsbienen und/oder Surrogatkönigin-nen, 5) oder 6) aus dem Gelege einer le-genden Arbeiterin kann eine Königin auf-gezogen werden. Weisel- und brutloseVölker können 7) ebenfalls einige Nach-kommen legender Arbeiterinnen aufzie-

hen, aber bald schrumpfen und schlie&szlig;lichausziehen. Schlie&szlig;lich 8) wird überhauptkeine Arbeiterbrut produziert und das Volkwird abnehmen und ausziehen (Abb 1). Ausgezogene Schwärme können einender brutlosen Entwicklungswege nehmen,oder sie vereinigen sich mit einem anderenausgezogenen Schwarm oder mit einem

niedergelassenen Volk. In solchen Fällenwerden überzählige Königinnen, legendeArbeiterinnen oder Ersatzköniginnen aus

dem neuen Volk entfernt.

Das Schicksal jeder einzelnen Königinhängt vom Brutstatus ab und von dem

Anteil und der Art der Arbeiterinnen-

Differenzierung, die nach dem Verlust derKönigin eintreten (Abb 2). Eine Differenzie-rung weiselloser Arbeiterinnen ist häufigund es gibt ein breites Spektrum ovariellerund pheromonaler Differenzierung (TabelleI). Das umfa&szlig;t sowohl Bienen, die nachden Pheromonen Arbeiter-ähnlich sind unddie: 1) Eier legen oder 2) dies nicht tun, alsauch Bienen, die Pheromone wie eine Kö-

nigin besitzen und 3) keine Eier legen oderdie 4) ein Gelege produzieren. Es gibtzahlreiche andere Zwischenstufen, die alsAbwandlungen der oben erwähnten Mög-lichkeiten entstanden sind. Sehr wenigeArbeitsbienen entwickeln sich zu echten

"Surrogatköniginnen", die in ihren Phero-monen sehr Königin-ähnlich sind und dieEier ablegen (Tabelle II).

Werden die relativen Häufigkeiten derOvar- und Pheromon-Differenzierung beiweisellosen Völkern nach den 8 Entwick-

lungswegen weiselloser Völker untersucht(Tabelle II, Abb 1), so scheint die Art derArbeiterinnen-Differenzierung durch kom-

plexe Interaktionen bestimmt zu werden,unter Beteiligung des sozialen Gefügesund der hierarchischen Konkurrenz zwi-

schen weisellosen Arbeiterinnen (Abb 2).

Apis mellifera capensis / Surrogat-Königin / legende Arbeitsbiene / Phero-mone

REFERENCES

Allsopp MH (1988) Mandibular gland acids andlaying workers in African honey bees. In:

Africanized Honey Bees and Bee Mites

(GR Needham, RE Page, M Delfinado-Baker,CE Bowman, eds) Wiley, New York, 72-79

Crewe RM (1988) Natural history of honey-beemandibular gland secretions: development ofanalytical techniques and the emergence ofcomplexity. In: Africanized Honey Bees andBee Mites (GR Needham, RE Page, M Delfi-

nado-Baker, CE Bowman, eds) Wiley, NewYork, pp 149-158

Free JB (1987) Pheromones of Social Bees.Chapman and Hall, London

Hemmling C, Koeniger N, Ruttner F (1979)Quantitative Bestimmung der 9-Oxodecensäure im Lebenzyklus der Kap-biene (Apis mellifera capensis Escholtz). Api-dologie 10, 227-240

Hepburn HR (1992) Pheromonal and ovarial de-velopment covary in Cape worker honey-bees, Apis mellifera capensis. Naturwissen-schaften 79, 523-524

Hepburn HR, Nefdt RJC, Whiffler LA (1988)Queen loss in the Cape honeybee: the inter-actions of brood, laying workers (falsequeens?) and queen cells. S Afr J Sci 84,778-780

Hepburn HR, Magnuson P, Herbert L, WhifflerLA (1991) The development of laying work-ers in field colonies of the Cape honeybee.J Apic Res 30, 13-16

Hillesheim E, Moritz RFA (1987) Genetic vari-ance of physiological characters in the Capehoneybee. J Apic Res 26, 30-36

Hillesheim E, Koeniger N, Moritz RFA (1989)Colony performance in honeybees (Apis mel-lifera capensis Esch) depends on the propor-tion of subordinate and dominant workers.Behav Ecol Sociobiol 24, 291-296

Jay SC (1968) Factors influencing ovary devel-opment of worker honeybees under naturalconditions. Can J Zool 46, 345-347

Jay SC (1970) The effects of various combina-tions of immature queens and worker beeson the ovary development of worker honey-

bees with and without queens. Can J Zool

48, 169-173Moritz RFA, Hillesheim E (1985) Inheritance of

dominance in honeybees (Apis mellifera ca-pensis Esch). Behav Ecol Sociobiol 17, 87-89

Ruttner F, Hesse B (1979) RassenspezifischeUnterschiede in Ovarentwicklung und Eiab-lage von weisellosen Arbeiterinnen der Ho-nigbiene Apis mellifica L. Apidologie 2, 159-183

Ruttner F, Koeniger N, Veith HJ (1976) Queensubstance bei eierlegenden Arbeiterinnender Honigbiene Apis mellifica L. Naturwis-senschaften 63, 434

Velthuis HW (1970) Ovarian development in

Apis mellifera worker bees. Entomol ExpAppl 13, 377-394

Velthuis HHW (1985) The honeybee queen andthe social organization of her colony.Fortschr Zool 31, 343-357

Velthuis HHW, Van der Kerk A (1988) Age, envi-ronment and genes in relation to the mandib-ular gland secretion of pure and hybrid Apismellifera capensis worker bees. In: African-ized Honey Bees and Bee Mites (GR Need-ham, RE Page, M Delfinado-Baker, CF Bow-man, eds) Wiley, New York, 79-86

Velthuis HHW, Ruttner F, Crewe RM (1990) Dif-ferentiation in reproductive physiology andbehaviour during the development of layingworker honey bees. In: Social Insects (W En-gels, ed) Springer, Berlin, 231-243

Whiffler LA, Hepburn HR (1991) Inhibition of

queen cell construction in the Cape honey-bee (Apis mellifera capensis). Apidologie 22,229-236