Testing Reliability of a Potential Island Mating Apiary Using DNA Microsatellites Peter Neumann, Job P

Home , Baltrum

Testing reliability of a potential island mating apiary using DNA microsatellites Peter Neumann, Job P. van Praagh, Robin F.A. Moritz, Jost H. Dustmann

To cite this version:

Peter Neumann, Job P. van Praagh, Robin F.A. Moritz, Jost H. Dustmann. Testing reliability of a potential island mating apiary using DNA microsatellites. Apidologie, Springer Verlag, 1999, 30 (4), pp.257-276. hal-00891583

HAL Id: hal-00891583 https://hal.archives-ouvertes.fr/hal-00891583 Submitted on 1 Jan 1999

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Original article

Testing reliability of a potential island mating apiary using DNA microsatellites

Peter Neumanna Job P. van Praagh Robin F.A. Moritza Jost H. Dustmann

a Martin-Luther-Universität Halle-Wittenberg, Fachgebiet Molekulare Ökologie, Institut für Zoologie, Kröllwitzerstr. 44, 06099 Halle/Saale, Germany b Department of Zoology and Entomology, Rhodes University, Grahamstown 61440, South Africa c Niedersächsisches Landesinstitut für Bienenkunde Celle, Wehlstr. 4a, 29223 Celle, Germany

(Received 25 July 1998; revised 21 November 1998; accepted 12 May 1999)

Abstract - Twenty-four virgin sister queens were kept for 21 days in mating nuclei on the drone-free island Baltrum to test the reliability of a potential mating area. On each of the neighbouring islands Nordemey and Langeoog (750 m and 2 km away) 12 sister queens were kept with drones. Workers from colonies with island-mated queens (Baltrum n = 11, Langeoog n = 7 and Nordemey n = 6) were genotyped with four DNA microsatellite loci (n = 996) to estimate queen mating frequency. No differences in queen mating frequency were observed between Langeoog and Nordemey. However, the level of polyandry on Baltrum was significantly lower than on the neighbouring islands, indicating that mating conditions were much more difficult. The standard genetic distance and differences in allele frequencies between the populations were determined to estimate putative origins of the drones. In this study, 43.7 % of the identified drone fathers did not descend from any of the queens on the adjacent islands. They were most likely from mainland colonies at least 5.4 km (3 km across open water) away, showing that the combination of distances over open water and over dry land is important in explaining the mating behaviour of honeybee queens. © Inra/DIB/AGIB/Elsevier, Paris Apis mellifera / DNA microsatellite / island / mating control / polyandry

1. INTRODUCTION around islands have a negative impact on the orientation of honeybee workers during Islands are routinely used as mating api- their flights [15]. Thus, islands have been aries to achieve controlled matings of virgin claimed to be ideal places because queens queens [34]. The large areas of open water and drones are not expected to cross open

* Correspondence and reprints E-mail: [email protected] water during their mating flights. However, derived from worker offspring and only bee breeders have repeatedly reported worker samples are needed to evaluate the uncontrolled matings even on these safe number of matings of the queen. The num- island mating areas. Recent studies of queen ber of queens which have mismated and the honeybee mating behaviour on drone-free number of times these queens mated with islands strongly support these observations unselected drones can be precisely deter- because they reveal that queens returned mined. from their nuptial flights with a mating sign even during high tide [33]. In this project we tested the reliability of a new on the island So far, the of potential mating apiary reliability mating apiaries of Baltrum has been tested with- (Germany) using virgin queens using virgin queens and DNA microsatellites. out drone colonies [5, 10, 17-19, 21, 27, 31], displacement experiments of drones [5, 11, 20] and marker phenotypes such as different races [2, 28] or mutants such as cor- 2. MATERIALS AND METHODS dovan [29]. Whereas some islands seem to provide controlled matings (e.g. [5, 20]) 2.1. Experimental design others apparently do not ([29] among others). A key factor is the distance between island and mainland. Successful mating Virgin sister queens (n = 48) were reared in summer 1995. Twenty-four of them were kept flights of virgin queens of more than 10 km in mating nuclei on the island of Baltrum which across water have never been open reported is free of other honeybee colonies. No foraging [10, 21, 30, 31]. Virgin queens were able workers could be observed before the experi- to cross at least 1 km or less across open ments and no drones could be attracted using a water [ 19, 29]. There are also reports on lure [14] during normal drone flight activity [25]. matings 7-8 km across open water [17, 18]. The distances from the Baltrum apiary towards Other authors [5, 20] could find no evi- the next available drone-producing colonies on the islands and Lan- dence of such long mating distances across neighbouring Norderney and on the mainland are in water for several North Sea islands. geoog (figure 1) given open table I. On the neighbouring island mating api- it seems as if distinct local charac- Thus, aries Langeoog and Norderney 12 queens each teristics of an island are also mating apiary were kept in the vicinity of 15 (Norderney) or important. 10 (Langeoog) drone-producing sister-queen In addition to the problems resulting from colonies (figure 1). No other bee apiaries are known on these islands. All queens were unusual test conditions [29] also the cor- virgin allowed to mate freely during a period of 21 days. dovan test may suffer from the pitfall of the Each queen was able to absolve mating flights marker with phenotypes interfering honey- at the age of 7 days onwards. The queens of the bee behaviour as shown for workers [13]. drone colonies on Norderney were daughters of The recent advance in honeybee DNA a single mother queen instrumentally insemi- microsatellite technology [8, 9] allows for a nated using mixed semen [22] of drones from genetical control of the reliability of mat- three sister-queen colonies. This results in a max- ing apiaries without interfering with honey- imum number of seven alleles per locus in the bee behaviour and routine bee breeding worker offspring on this island. On Langeoog a maximum number of 12 alleles per locus was practice. DNA microsatellites can be used to possible. Sealed worker brood samples (n = 50 assess the number of in precisely patrilines per queen) were taken from colonies with mated a DNA tests can be honeybee colony [9]. queens and raised isolated in an incubator. easily incorporated in the routine procedure Recently emerged workers were immediately at mating apiaries. The genotypes of the stored in 96 % ethanol at -15 °C until DNA mother queen and her drone mates can be extraction. We used four DNA microsatellites which were developed by Estoup et al. [8, 9]. Multi- plex PCR was performed using two pairs of loci (A43/B 124, A76/A 107) and the standard proto- cols given in Estoup et al. [8, 9]. Amplification products were electrophoresed on 6 % poly- acrylamide sequencing gels for 5.5 h (A76/A107) or 5 h (A43/B 124) with M13mp18 control DNA sequencing reactions run on the same gel as size standards. Microsatellite alleles were scored as fragment lengths in base pairs.

2.3. Genotype analysis 2.2. DNA isolation and number of matings and microsatellite analysis The genotypes of the mother queens and the DNA was phenol extracted from single work- father drones were derived from the genotypes of ers (n = 40 per colony) following routine proto- the sampled workers. The queen was assumed cols [1] with the following changes: to be homozygous when an allele was present in worker of the The was con- 1) workers were incubated with agitation in every colony. queen insect Ringer solution (127 mM NaCl, 1.5 mM sidered to be heterozygous when every worker CaCl2, 5 mM KCl, pH 7.4 with NaOH) for 1 h at carried one of two alleles. The paternal alleles RT before extraction; were those not carried by the queen. We used the putative genotype of the mother queen to 2) single worker thoraces were homogenised exclude additional allele combinations. If multi- in 400 of DNA extraction buffer mM μL (100 ple queen genotypes were possible at a given 100 mM Tris-HCl 10 mM NaCl, (pH 8.0), NaCl, locus we chose, as a rule, the allele combination 0.1 % SDS); yielding the lowest number of observed matings 3) DNA was resuspended in 30 μL DDH2O. (n0). In case a drone’s genotype could not be unambiguously determined owing to heterozy- bility that two genes, one drawn randomly from gosity of the queen at that locus, we assumed an population 1 and the other from population 2, equal possibility for yielding one or the other are identical. This set was calculated for each of queen allele for calculating the allele frequen- the four loci. Then, the average for all loci was cies of the drone populations. calculated in each of the three cases: (J1, J2, J12).

2.4.4. 2.4. Data Improved Bonferroni analysis procedure for χ2-tests 2.4.1. Number of estimated matings (k) We calculated a χ2-test for each allele at each locus to evaluate if the alleles shown the As a result of finite sample sizes the number by drones which had mated with the queens on Bal- of observed patrilines may severely underesti- and have a common mate the actual number of subfamilies. There- trum, Norderney Langeoog allele pool or not. Since the high number of alle- fore, we estimated the number of patrilines in an les at the used microsatellite loci cause infinite sample following Comuet and Aries [4]: may significant differences only by chance, we used an improved Bonferroni procedure [16, 32] to adjust the significance levels. We also used this procedure to combine the dependent test results because the test statistic cannot be split up into where is the number of E(k) expected patrilines independent test statistics [16, 32]. For n test in the k is the number of colony, equally frequent and as a statistics, q1, q2, ..., qn for Qi continu- patrilines and n is the sample size. ously distributed statistic for testing the null We followed Oldroyd et al. [26] and numer- hypothesis H0,i versus the alternative hypothe- ically evaluated k by substituting E(k) with our sis H1,i (i = 1, ....., n) the overall hypothesis Hs is observed number of matings (no) and the worker rejected if for at least one i: sample sizes for n. 2.4.2. Number of effective males (me) where p(i) are the ordered p-values for χ2-tests, α(i) is the significance level for the subhypothe- The average intracolonial relatedness G was estimated according to Estoup et al. [9]. Then, sisi. H the number of effective males (me) was calcu- For each Hi α(i) was calculated as follows: lated following Chevalet and Cornuet [3]:

with α = 0.05. where me is the number of effective males and G is the average intracolonial relatedness. 2.5. Putative origin of Baltrum worker bees’ fathers 2.4.3. Genetic distance The genotypes of all drones which mated with We used the standard genetic distance of [23]: the tested Baltrum queens were compared with the genotypes of the drones which mated with the queens from Langeoog and Norderney. Bal- trum drones showing allele combinations that did not correspond with the drone genotypes of one island were with excluded from that potential source. Baltrum drone fathers which might potentially originate from the neighbouring island mating apiaries Langeoog or Norderney were determined and the differences in allele where is the that two frequencies J1 probability randomly towards the drones which mated with the Lan- chosen genes in population 1 are identical, is J2 geoog and Norderney queens were evaluated the same for and is the population 2, J12 proba- using the Bonferroni procedure. 2.6. Comparisons of queen mating tion of the locus B124 in the case of Bal- frequencies trum and Norderney significant differences for the allele frequencies at all loci. For We computed Mann-Whitney U-tests to esti- Norderney χ2-tests were not calculated at mate potential differences for the number of the locus A76 because drones could be def- observed and estimated queen matings and for initely excluded owing to specific alleles. the effective number of males between the three Thus, we could exclude 43.7 % of the Bal- islands. To determine the probability of identical trum drones from any of the drone colonies drone genotypes we estimated for each apiary the product of the highest allele frequency for on the adjacent islands. They most likely each locus. came from mainland colonies which were at least 5.4 km away. We found that 50.7 % of the drones could have originated from 3. RESULTS the neighbouring island mating apiary Lan- geoog since they had genotypes in common with the drones. A total number of 996 workers was geno- corresponding However, these drones from typed and assigned to patrilines (table II). potentially originating showed differences Seventy-one patrilines were observed on Langeoog, significant in the allele of the drones which Baltrum (table II). The ranges and the mean frequencies numbers (x ± SE.) of the observed and estimated with the queens on Langeoog (table Four drones which had mated mated queen matings and of the effective V). (5.6 %) number of males for all three islands are with the Baltrum queens showed alleles common to both given in table III. We found a mean number neighbouring mating apiaries and could not be excluded from of 6.45 ± 1.27 observed queen matings on any source. Baltrum. For Langeoog we found a mean of 12.5 ± 1.23 observed paternities. The The standard genetic distances between mean number of observed was 12.8 matings the tested populations are shown in table VI. ± 2.27 on For two colonies 1 Nordemey. (L High distances were observed between the and the estimated numbers of N5) matings drones mated with the queens from Norder- were than the observed substantially higher ney and the Baltrum patrilines and between number of patrilines, which can be attributed the drones from Norderney and Langeoog to low sizes. these sample Therefore, whose drone mothers originated from unre- colonies were excluded from the compar- lated breeding lines. This also shows that isons of For the queen mating frequencies. matings of the Baltrum queens with Norder- other colonies the number of estimated mat- ney drones were most unlikely. An inter- ings was slightly higher than the observed mediate distance was observed between the number of patrilines, showing that sample drones from Baltrum and Langeoog. As sizes were We found sufficiently large. sig- expected, low distances were found between nificant for the number of differences the tested sister queens. observed and estimated queen matings and for the number of effective males between Baltrum and the islands neighbouring (Mann 4. DISCUSSION Whitney U-test: Baltrum/Langeoog P < 0.01; P < 0.05). Baltrum/Norderney Our results clearly show that successful However, we failed to detect dif- significant took on the drone-free ferences between and mating flights place Langeoog Norderney island of Baltrum the next available (Mann U-test: P > 0.05). The allele although Whitney source of mature drones was at frequencies for all tested microsatellite loci sexually least 5.4 km away. are given in table IV. Our results of the χ2-tests are shown in table V. We found for The Baltrum queens showed a signifi- all tested drone populations with the excep- cantly smaller number of matings compared

to the queens which were mated on the crossed the distance to the next available neighbouring island mating apiaries. Poten- drone source while the mud flats around tial differences in honeybee queen mating Baltrum fall dry at low tide. In light of the frequency which may be attributed to dif- observations of Van Praagh et al. [33] this ferent types of mating apiaries [24] or to does not seem to play a role because queens genetic variability among honeybee races returned from their nuptial flight with a mat- (Neumann, unpublished data) can be ing sign even when the tide was high at Bal- excluded. Thus, the number of queen mat- trum. ings on Baltrum certainly depended upon the drone-free conditions, showing that mat- The allele combinations of the drones with the Baltrum ing conditions were much more difficult which mated queens than under normal beekeeping practice. enabled us to exclude 43.7 % of them from any of the used queens. Nine of eleven The majority of the tested Baltrum mated Baltrum queens certainly interacted queens (81.8 %) mated with males which with other drone sources probably from the most did not from the likely originate neigh- mainland. An interaction with Nordemey island bouring mating apiaries Langeoog is because we could exclude the and These drones unlikely Norderney. probably majority of the Baltrum patrilines from that derived from colonies on the mainland more (94.4 %). Furthermore, the than 3 km origin potential away. ’Norderney’ drones which mated with the This is the second largest distance ever Baltrum queens showed alleles (either 127 bp reported for successful honeybee mating for locus A43 or 291 bp for locus A76) flights across open water after those of Klatt which were very rare in the patrilines on [17, 18]. However, Evenius [12] doubted Norderney. In the case of the potential the drone-free conditions on the peninsula Norderney drones, interactions with at least Frisches Haff during that time. In our exper- two different drone sources (Norderney and iment these drones could also potentially mainland or Norderney, Langeoog and originate from undetected swarms on Bal- mainland) must have occurred. Given that trum. However, no drones could be attracted queens were searching for drones this seems using a lure during normal drone flight activ- to be most unlikely. It cannot be ruled out ity [7]. Although we cannot definitely rule that 50.7 % of the drones might originate out that drones remained undetected in the from Langeoog which is 1.7 km over open lure experiments we consider it unlikely that water away. Two queens showed only the Baltrum queens have mated with drones potential Langeoog progeny in their worker from that island. It seems more likely that offspring. However, we found significantly the queens fly to the mainland for mating. different allele frequencies between these One could argue that the queens might have potential ’Langeoog’ drones and the drones

which mated with the Langeoog queens. the potential mating area Baltrum with the We therefore reject the hypothesis that the current state of evidence. We recommend drones originated from the same gene pool. a critical testing of island mating stations Following our argumentation given for less than 8 km apart from the mainland [17, Norderney we consider it improbable that 18]. the queens had visited two possible drone sources their The stan- during mating flights. ACKNOWLEDGEMENTS dard genetic distances also indicate that matings with drones from the neighbouring We are grateful to V. Schwarz for technical island Norderney were unlikely. assistance. We wish to thank F.K. Böttcher for This seems surprising in light of the small help during literature search. G. Koeniger, A. Stokes and two referees distance over open water between Norder- anonymous provided valuable comments on the manuscript. Financial ney and Baltrum. Similarly, matings of the support was granted by the DFG. Baltrum queens with drones from the mainland were clearly indicated in spite of the largest distance over open water. These Résumé - Fiabilité d’un rucher de fécon- results show that a combination of distances dation éventuel testée à l’aide de micro- over water and over land is open dry impor- satellites d’ADN. Vingt-quatre reines soeurs tant to the observed behav- explain mating ont été placées durant 21 j sur l’île de Bal- iour of orientation queens. Queen flights trum, qui est dépourvue de mâles, afin de during low tide [33], local wind directions, tester la fiabilité d’une éventuelle station de the southwest position of the sun during fécondation. Sur chacune des îles voisines, time in the afternoon queen mating flight Langeoog et Norderney, il y avait 12 autres and more attractive visual cues potentially reines soeurs avec des colonies de mâles provided by the mainland may also be (figure 1). Le tableau I donne les distances important. des colonies de Baltrum aux plus proches Our results show that queens most likely ruchers producteurs de mâles. On a prélevé have the potential to successfully mate with des cadres de couvain d’ouvrières dans cha- mainland drones. Clearly, we cannot give cune des colonies possédant une reine fécon- a final judgement of the reliability of Bal- dée (n = 11 à Baltrum ; n = 7 à Langeoog et trum as an established mating apiary. A reli- n = 6 à Norderney). On a établi le génotype ability testing is only possible under nor- de 996 ouvrières fraîchement écloses en uti- mal beekeeping conditions with a sufficient lisant quatre locus différents de microsatel- number of drone colonies [6, 28]. We con- lites d’ADN pour estimer la fréquence sider it unlikely that queens would mismate observée et la fréquence estimée d’accou- if an adequate number of drones is avail- plement des reines, ainsi que le nombre able on Baltrum. Nevertheless, absolutely effectif de mâles (tableaux II et III), et pour controlled matings such as with instrumen- obtenir l’origine supposée des mâles aux- tal insemination cannot be guaranteed on quels les reines de Baltrum s’étaient accou- plées. On a recherché entre les populations comportement d’accouplement des reines des différences éventuelles dans les fré- de Baltrum de combiner les distances à par- quences alléliques à l’aide du test Chi2 courir au-dessus de la mer et au-dessus de la (tableau IV) et de la méthode Bonferroni terre ferme jusqu’aux colonies de mâles les (tableau V) et calculé les distances géné- plus proches (tableau I). tiques selon Nei (tableau VI). Le problème reste ouvert de savoir si des Les fréquences d’accouplement sur Baltrum reines qui s’accouplent dans des conditions i.e. avec un nombre satisfaisant (n = 71 accouplements) sont significative- normales, de sur une station de ment plus faibles que sur les stations des de colonies mâles, îles voisines (tableaux II et III, test U de fécondation établie à Baltrum, s’accouplent avec des mâles non Mann-Whitney : Baltrum/Langeoogp > 0,01 ; aussi sélectionnés, l’installation de reines sans Baltrum/Norderney p < 0,05). On n’a pas puisque vierges trouvé de différence dans la fréquence colonies de mâles ne convient pas en condi- la fiabilité d’accouplement entre les îles Langeoog et tions de routine pour évaluer d’une station de fécondation. Nos résultats Norderney, ce qui laisse penser que les montrent de claire des conditions d’accouplement sur l’île de Bal- pourtant façon que contrôlés ne être trum sont plus difficiles. Les mâles qui accouplements peuvent pas sur l’île de Baltrum. Soit les mâles, s’étaient accouplés avec les reines de Bal- garantis soit les reines, soit les deux sexes sont trum présentaient des combinaisons d’allèles de franchir de distances qui ne correspondaient pas aux génotypes capables grandes mer vols de des mâles des îles voisines. 43,7 % des mâles au-dessus de la lors de leurs fécondation. Nous recommandons donc de identifiés ne pas des colonies provenaient la fiabilité des stations de fécondation de mâles des îles voisines. Ils venaient selon tester situées sur des îles à moins de 8 km du conti- toute vraisemblance du continent; ils avaient nent. © Inra/DIB/AGIB/Elsevier, Paris donc survolé la mer sur une distance d’au moins 5,4 km (tableau I). On ne peut pas Apis mellifera / accouplement controlé / exclure que 50,7 % des mâles provenaient de polyandrie / île /microsatellite Langeoog. Ces mâles présentaient néan- moins des fréquences alléliques significativement différentes de celles des mâles qui Zusammenfassung - Prüfung der Paa- s’étaient accouplés avec les reines de Lan- rungssicherheit einer potentiellen Insel- geoog (tableau V). Quatre mâles avaient des belegstelle mit DNA-Microsatelliten. Vier- allèles présents dans les deux stations de undzwanzig Geschwisterköniginnen wur- fécondation et aucune origine ne pouvait den für 21 Tage auf der drohnenfreien Insel être exclue pour eux. Les populations de Baltrum plaziert, um die Zuverlässigkeit mâles de Langeoog et de Norderney et les einer potentiellen Belegstelle zu testen. Auf lignées paternelles de Baltrum ont présenté den benachbarten Inseln Langeoog und Nor- des fréquences alléliques significativement demey standen je 12 weitere Geschwisterkö- différentes (tableau V). Cela signifie que niginnen mit Drohnenvölkem (Abbildung 1). les accouplements entre les mâles de Lan- Die Abstände von den Baltrum Völkern zu geoog ou de Norderney et les reines de Bal- den nächsten Drohnen produzierenden Bie- trum sont improbables. On a trouvé une dis- nenständen sind Tabelle I zu entnehmen. tance génétique élevée entre les mâles de Aus jedem Volk mit einer begatteten Köni- Norderney et de Baltrum (tableau VI), ce gin (n = 11 Baltrum; n = 7 Langeoog und qui montre aussi que les accouplements avec n = 6 Norderney) wurden Arbeiterinnen- des mâles de Norderney sont très impro- brutwaben entnommen. Isolierte, frisch bables, bien que la distance à parcourir au- geschlüpfte Arbeiterinnen (n = 996) wur- dessus de la mer soit plus faible. Ces résul- den unter Verwendung vier verschiedener tats montrent qu’il est important pour le DNA Microsatelliten Loci genotypisiert (Tabelle II), um die beobachtete und zwischen den Drohnen von Norderney und geschätzte Paarungshäufigkeit der Köni- Baltrum gefunden (Tabelle VI), was eben- ginnen sowie die effektive Anzahl an Männ- falls zeigt, daß Paarungen mit Drohnen von chen abzuschätzen (Tabellen II und III) und Norderney am unwahrscheinlichsten sind, um die vermutliche Herkunft der Drohnen obwohl die Entfernung über offenes Wasser mit denen sich die Baltrum Königinnen am geringsten ist. Diese Ergebnisse deuten gepaart haben zu ermitteln. Zwischen den daraufhin, daß eine Kombination aus Ent- getesteten Populationen wurden mögliche fernungen über Wasser und über Land zu Unterschiede in den Allelfrequenzen den nächsten Drohnenvölkern für das Paa- (Tabelle IV) mit Hilfe von χ2-Tests und der rungsverhalten der Baltrum Königinnen von Bonferroni Methode untersucht (Tabelle V) Bedeutung ist (Tabelle I). Ob sich Köni- und die genetischen Distanzen nach Nei ginnen unter regulären Bedingungen auf berechnet (Tabelle VI). Die Paarungshäufig- einer etablierten Belegstelle Baltrum, d.h. mit einer ausreichenden Anzahl von Droh- keiten auf Baltrum (n = 71 Paarungen) waren signifikant geringer als auf den benachbar- nenvölkem, auch mit unselektierten Droh- ten Inselbelegstellen (Tabellen II, III, Mann nen paaren, bleibt offen, da das Aufstellen ohne Drohnen- Whitney U-test: Baltrum/Langeoog p < 0,01; unbegatteter Königinnen völker nicht für die der Zuver- Baltrum/Norderney p < 0,05). Es konnten Einschätzung keine Unterschiede in der Paarungshäufig- lässigkeit einer Belegstelle unter Routine- keit zwischen Langeoog und Norderney bedingungen geeignet ist. Unsere Ergebnisse gefunden werden (Mann Whitney U-test: demonstrieren jedoch eindeutig, daß auf Baltrum kontrollierte nicht p > 0,05), was darauf hindeutet, daß die Paa- Paarungen garan- rungsbedingungen auf Baltrum erschwert tiert werden können. Entweder Drohnen sind. Die Drohnen, mit denen sich die oder Königinnen oder beide Geschlechter sind in der bei ihren Baltrum Königinnen gepaart haben, konnten Lage, Paarungsflügen Strecken offenen Wassers zu über- ihrer Allele nicht von den Nach- größere aufgrund Wir daher eine Über- barinseln stammen. 43,7 % der identifizier- queren. empfehlen der Sicherheit von ten Drohnen stammen nicht von den Droh- prüfung Inselbelegstellen, die als 8 km vom Festland entfernt nenvölkern der benachbarten Inseln. Sie weniger liegen. © Inra/DIB/AGIB/Elsevier, Paris kamen am wahrscheinlichsten vom Fest- das am weitesten über offenes Was- land, Apis mellifera / Insel / DNA-Microsatelliten / ser entfernt Es kann nicht lag (Tabelle I). Paarungskontrolle / Polyandrie ausgeschlossen werden, daß 50,7 % der Drohnen von Langeoog stammen. Diese Drohnen unter- zeigten jedoch signifikant REFERENCES schiedliche Allelfrequenzen zu den Droh- nen, mit denen sich die Langeoog Königin- [1] Beye M., Raeder U., Rapid DNA preparation nen gepaart haben (Tabelle V). Vier from bees and % GC fractionation, Bio Tech- Drohnen wiesen auf beiden Belegstellen niques, 14 (1993) 372-374. vorkommende Allele auf und konnten von [2] Böttcher F.K., Was taugt die Belegstelle?, Der Imkerfreund 2 115-116. keiner Herkunft ausgeschlossen werden. (1947) unterschiedliche [3] Chevalet C., Comuet J.M., Etude théorique sur la Signifikant Allelfrequen- sélection du caractère « production du miel » zen (Tabelle V) zeigten die Drohnen Popu- chez l’abeille. I . Modèle génétique et statistique, lationen von Langeoog und Norderney und Apidologie 13 (1982) 39-65. die Patrilinien von Baltrum. Dies deutet dar- [4] Cornuet J.M., Aries F., Number of sex alleles in aufhin, zwischen Drohnen a sample of honeybee colonies, Apidologie 11 daß Paarungen (1980) 87-93. von Langeoog oder Norderney und den [5] Drescher W., Untersuchungen zur Zuflugsicher- von Baltrum unwahrscheinlich Königinnen heit der Inselbelegstelle Mellum, Z. Bienenforsch sind. Eine hohe genetische Distanz wurde 8 (3) (1965) 49-53. [6] Drescher W., Betrachtungen zum Thema Beleg- [20] Meinen G., Wie sicher ist die Inselbelegstelle stellen, Allg. Dtsch. Imkerztg. (8) (1974) 57-62. Spiekeoog vor fremdrassigem Drohnenzuflug?, NW Dtsch. 22 128-130. [7] Dustmann J.H., van Praagh J.P., Faltus E., Imkerztg. (1970) Maarzahl H., Die Insel Baltrum als Belegstelle?, [21] Minderhoud M., Verslag der proeven over konig- Dtsch. Bienen. J. 4 (6) (1997) 266. innenentelt op Urk, Maandschrift voor Bijeen- teelt 26 40. [8] Estoup A., Solignac M., Harry M., Cornuet J.M., (3) (1923) Characterisation of (GT)n and (CT)n microsatel- [22] Moritz R.F.A., Homogenous mixing of honey- lites in two insect species: Apis mellifera and bee semen by centrifugation, J. Apic. Res. 22 Bombus terrestris, Nucleic Acids Res. 21 (1993) (1983) 249-255 1427-1431. [23] Nei M., Molecular Evolutionary Genetics, [9] Estoup A., Solignac M., Cornuet J.M., Precise Columbia University Press, New York, 1987. assessment of the number of patrilines and [24] Neumann P., Moritz R.F.A., van Praagh J., Queen genetic relatedness in honeybee colonies, Proc. R. mating-frequency in different types of honeybee Soc. Lond. B 258 (1994) 1-7. mating apiaries, J. Apic. Res. 38 (1999) in press. [10] Evenius J., Königinnenzuchtversuche auf der [25] Oertel E., Observations on the flight of drone Greifswalder Oie, Leipz. Bienenztg. 44 (12) honey bees, Ann. Entomol. Soc. Am. 49 (1956) (1929) 244. 497-500. [11] Evenius J., Zur Frage der Belegstellen, Westf. [26] Oldroyd P.B., Clifton M.J., Wongsiri S., Rinderer Bienenztg. 45 (1930) 293-294. T.E., Sylvester H.A., Crozier R.H., Polyandry in the Evenius Die neuzeitliche Bienenzucht und die genus Apis, particularly Apis andreniformis, [12] J., Behav. Ecol. Sociobiol. 40 (1997) 17-26. Verwendung von Inseln als Belegstellen, Bie- nenwirtsch. Zentrbl. (4) (1931) 92-95. [27] Oortman Gerling E., Betrouwbaarheids-proef op Marken 1992, Buckfast Bull. 15 (1993) 31-32. [13] Frumhoff P.C., The effect of the cordovan marker Ruttner F., und Drohnendichte, on kin discrimination nestmate [28] Drohnenflugweite apparent among Dtsch. Bienenwirtsch. 42-47. honeybees, Anim. Behav. 42 (1991) 854-856. 10 (1) (1959) Ruttner Ruttner Wie weit Drohnen Observations of behaviour in [29] F., H., fliegen [14] Gary N.E., mating und Bienenvater 86 the Res. 2 3-13. Königinnen?, (1) (1965) honeybee, J. Apic. (1963) 15-21. [15] Heran H., Lindauer M., und Windkompensation [30] Sladen F.W.L., Why young queens may fail, Seitenwindkorrektur bei Flug über Wasser, Glean. Bee Cult. 48 (2) (1920) 50-52. Z. Vgl. Physiol.47 (39) (1963) 39-55. [31] Sladen F.W.L., Queen mating experiments, [16] Hochberg Y., A sharper Bonferroni procedure Ebenda 48 (12) (1920) 717-718. for tests Biometrics 75 multiple of significance, [32] Simes R.J., An Bonferroni (1988) 800-802. improved procedure for multiple tests of significance, Biometrics 73 [17] Klatt G., Züchtungsmöglichkeiten an der (1986) 751-754. X Wasserkante, Arch. Bienenkd. 8 (1929) [33] Van Praagh J., Neumann P., Moritz R.F.A., 318-321. Koeniger G. and Dustmann J.H., Mating [18] Klatt G., Unsere ostpreußischen Belegstellen, behaviour of honeybee queens (Apis melli- Die deutsche Bienenzucht in Theorie und Praxis fera L.) on an island free of drones, Pszcz. Zesz. 37 (3) (1929) 74-77. Nauk. XL 2 (1997) 25-32 [19] Klöpping B., Hompelvoet is getest, Buckfast [34] Zander E., Böttcher F.K., Haltung und Zucht bei Bull. 16 (1993) 15. der Biene, Eugen Ulmer Verlag, Stuttgart, 1989.