Position of the Red bee, (Buttel-Reepen 1906), within the Genus Apis F. Ruttner, D. Kauhausen, N. Koeniger

To cite this version:

F. Ruttner, D. Kauhausen, N. Koeniger. Position of the Red , Apis koschevnikovi (Buttel- Reepen 1906), within the Genus Apis. Apidologie, Springer Verlag, 1989, 20 (5), pp.395-404. ￿hal- 00890794￿

HAL Id: hal-00890794 https://hal.archives-ouvertes.fr/hal-00890794 Submitted on 1 Jan 1989

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

Position of the Red Honey bee, Apis koschevnikovi (Buttel-Reepen 1906), within the Genus Apis

F. Ruttner D.2 Kauhausen N. Koeniger

1 Universität Frankfurt, Institut für Bienenkunde (Polytechn. Gesellschaft), D-6370 Oberursel, FRG 2 Bayerische Landesanstalt für Bienenzucht, D-8520 Erlangen, FRG (received 19 December 1989, accepted 29 March 1989)

Summary — The "Red Bee" of , described in 1988 by Koeniger et al. and Tingek et al. as a separate species, was first named by H. v. Buttel-Reepen in 1906. The correct name, therefore, is Apis koschevnikovi Buttel-Reepen 1906 and not Apis vechti Maa 1953. By multivariate analysis (PCA and DA) Apis koschevnikovi can clearly be separated from the 2 other cavity-nesting Apis spe- cies. Although A. koschevnikovi is very similar to A. cerana phenotypically, it is much larger than bees of this species from a similar geographic latitude, corresponding in size to equatorial popula- tions of A. mellifera from Africa.

Apis— — Apis koschevnikovi - morphometrics — Borneo

THE CORRECT SCIENTIFIC NAME of his new form, but elsewhere in the same work he referred to it as koschevni- no koschev- In the spring of 1988, Koeniger et al. and kovi. There is doubt but that nikovi itself is the name for this Tingek et al. published data on the taxo- by bee]. To him, the at the Berlin Mu- nomic status of a special honey bee from specimens seum were evidence that A. cerana also S.E. Asia as a true species which they de- occurs on the African continent. termined as Apis vechti (Maa, 1953). But this was not the first description of this About 50 years later (1953) Maa de- honey bee, very conspicuous by its red- scribed a number of specimens of the dish legs and hair, and by its size. Buttel- "Red Bee", all collected in Borneo, in his Reepen (1906) found several specimens taxonomic revision of the genus Apis. of this kind in the Museum of Natural His- Maa considered the "Red Bee" to be a tory in Berlin which were labelled as separate species and gave it a new "Cameroon", and one as "N. Borneo". He name, Apis vechfi (after the entomologist dismissed the possibility as unlikely, and J. Van der Vecht then of the Bogor Mu- named this cerana-type honey bee "Apis seum, now Putten, The Netherlands). He mellifica indica-koschevnikovi " [in Buttel- even noticed a certain geographic vari- Reepen’s taxonomic system A. indica (= ability sufficient to establish 2 subspecies " A. cerana) was taken as a subspecies of ("vechti and "linda"). Maa did not re- A. mellifera. The prefix "indica" was used evaluate the specimens of the Berlin Mu- by Buttel-Reepen to show the relationship seum and proposed retaining the name &dquo;A. koschevnikovi &dquo; for the supposedly Af- formis, all clearly different from A. koschev- rican &dquo;Red Bee&dquo;. nikovi. These findings raise the question of the of this between the The present situation is as follows : position species other 4 known recent Apis species. 1) &dquo;Apis vechti &dquo; is a true species: it differs from A. cerana in the endophallus (Tingek et al., 1988) and time of flight (= re- productive isolation; Koeniger et aL, 1988). It is with A. cerana indica. sympatric MATERIAL AND METHODS 2) There is no evidence yet that a cerana- type honey bee exists on the African conti- For the of the &dquo;Red Bee&dquo;, nent. of the at the morphometric analysis Mislabelling specimens two sets of data were used. Berlin Museum is the most probable expla- nation for the present confusing situation. Consequently, 2 names are probably used for the same bee of the same area. After Phenetic analysis including fossil hon- consulting Dr Charles Michener in this eybees matter, we asked Dr Frank Koch, Museum fur Naturkunde Universität, Ber- (Humboldt Characters of 12 individuals of A. koschevnikovi to re-examine the de- lin) specimens and 12 of A. cerana indica, both from the region scribed in 1906 by Buttel-Reepen (which of Tenom, , N. Borneo, were analysed to- have miraculously survived the last 83 gether with data on individual honey bees as years) to confirm their cerana-type charac- used in an earlier analysis (Ruttner et al., 1986). ters. The following criteria were used The origin of these bees was the following : which discriminate best between A. cerana 1) A. armbrusteri— Schwabische Alb, Upper Mi- and A. mellifera : ocene, 10 specimens i) prolongation of radial vein of hind wing 2) A. dorsata - , 10 specimens (&dquo;indica vein&dquo;) - present; 3) A. florea- Pakistan, 10 specimens ii) tomentum on last tergite-present; 4) A. cerana cerana - , 10 speci- mens iii) number of hamuli on hind wing - 17 to 19. 5) A. mellifera— Iran, 10 specimens. The characters used were 16 venation Thus there is no doubt that all of Buttel- wing angles; no characters of size were included. The are Reepen’s specimens cerana-type and multivariate statistical methods were factor and that they were not collected in Cameroon discriminant analyses (PCA, DA). Both methods but most likely in Borneo. According to the gave basically identical results. Finally, ellipses rules of priority, the name &dquo;A. vechti &dquo; is re- of confidence were calculated (Cornuet, 1982). dundant and the correct scientific name of the Red Honey bee has to be Apis ko- schevnikovi (Buttel-Reepen, 1906). Taxonomic-morphomefric analysis Here it should be noted that no type similar to the &dquo;Red Bee&dquo; is among the vari- In these all 34 characters were used ous honeybees collected by Wallace in analyses as in the standard method of bee mor- North Borneo and described Smith honey by phometry (Ruttner et al., 1978; Ruttner, 1988); he lists A. dorsata and a va- (1858; 1865); not individual bees but the means of samples of riety of this species, A. testacea, A. indica 15-20 bees were taken as units. Five samples with the variety perrottetii and A. andreni- of A. koschevnikovi from N. Borneo and 1 sam- ple from were analysed together with cluster, with only slight differentiation of samples of A. cerana of various origins and of the 2 species (Figs. 1 and 2). A. koschev- tropical populations (Arabia, E. Africa) of A. mel- nikovi with the 2 other lifera. appears together cavity-nesting species as 3 separate clus- Voucher of A. koschevnikovi have specimens ters without A. cerana in the been for in the overlapping, placed permanent preservation centre and A. mellifera at the Honeybee Taxonomy Collection of the Institut periphery. fur Bienenkunde Oberursel, University of Frank- The cluster of A. koschevnikovi is found in furt. direct contact with that of A. cerana, but distant from the mellifera cluster. The cera- na samples from Borneo are at the centre of the general cerana cluster (Fig. 2B). In the graph showing the ellipses of confi- RESULTS dence (95%, Fig. 2), the K cluster heavily overlaps the C cluster but not the M clus- ter. Phenetic analysis

As was to be the of the expected, positions Taxonomic analysis clusters are identical with those of an earli- er investigation of fossil and known recent species : the fossil A. armbrusteri and A. The samples of A. cerana are arranged in dorsata are represented as a common 2 separate groups (Fig. 3). While the popu- lations with small bees of the subspecies tween the centroids of A.c. cerana and A. A. cerana indica from S. , , c. indica. , , etc. appear to the left, If the group A. koschevnikovi-A. cerana the of the A. cerana cerana samples large is itself, that is exclu- from Afghanistan, Pakistan, N. India, Chi- analysed by including sively samples of these 2 species, the ce- na, and those of A. cerana japonica are to rana in one the right. The clusters of the tropical melli- subspecies appear compact, only subdivided cluster, while the fera populations are at great distance in slightly koschevnikovi are found in a re- the far right corner. samples mote, compact aggregation (Fig. 5). The cluster of the 5 A. koschevnikovi One 20 from samples (15 bees each) from Borneo is sample (No. 991; bees) Sumatra, collected in 1978 by one of the found between these 2 groups (Fig. 3), authors (Koeniger) in Muaro near Solok, is close to that of A. cerana cerana, the clos- separated from the A, c. indica cluster by a est samples the populations from being great distance, but very close to the A. S.W. (, Fig. 3). When ellip- koschevnikovi samples (Fig. 3). These ses of confidence are calculated, A. kos- honey bees show the same typical rufous chevnikovi with A. c. in overlaps japonica color of legs and abdomen as A. koschev- the graph of factor 1/factor 2, but it is well nikovi; their size is somewhat smaller than separated in the graphs of factors 1/3 (Fig. that of A. koschevnikovi from Borneo, but 4). In this analysis (1/2) the distance of the they are much larger than indica bees from A. koschevnikovi centroid to the centroid of Sumatra, and the cubital index and the in- A. c. cerana is less than the distance be- dex of st6 show the typical extreme values of A. koschevnikovi. The following data are cerana if very different, remote species are arranged in the sequence A. koschevniko- included such as the Miocenic A. armbrus- vi Borneo — sample 991 (Sumatra) - A. teri, a fossil honeybee which lived 12 mil- c. indica Sumatra (6 samples; in mm) : lion years ago (Fig. 2). But even in this length of fore wing, 8.541-8.264-7.769; case, almost no overlapping occurs with length of hind leg, 7.612-7.558-6.839; ter- the ellipse of confidence (95%) of sympa- gites 3 + 4, 3.987-3.886-3.535; index of tric A. c. indica from Borneo. A good dis- sternite 6, 89.56-90.24-86.50; cubital in- crimination is obtained, however, if only dex, 7.59-7.86-4.34. There is no doubt, the closely related cavity-nesting species therefore, that the sample from Sumatra A. cerana and A. mellifera are included in belongs to the same taxonomic unit as the the analysis (Figs. 3, 4 and 5). of A. koschevnikovi from Borneo. samples In A. koschevnikovi, only a few of the characters generally used in honey bee morphometrics are outside or at an ex- treme end of the known range of A. cerana Ruttner, A. koschevnikovi is DISCUSSION (see 1988). very long-tongued (5.870 mm), slender (st6-1 89.6), with narrow tomenta (tom-1 Apis koschevnikovi can be separated from 0.389) and high CI (7.59). All the other all the other Apis species by quantitative characters, especially those of size, are characters (Figs. 1, 3, 5), but the level of well within the range of cerana populations confidence is low for discrimination from A. of the temperate zone. The most conspicu- ous character by which &dquo;this species can geographic latitudes. The samples of the very easily be distinguished from any other &dquo;Red Bee&dquo; were collected in N. Borneo honey bee species is the uniform rufous (Sabah) at 2° N latitude. The sympatric A. pattern&dquo; (Maa, 1953). cerana subsp. indica is much smaller (Rin- Evidently, the overlapping of the A. ko- derer et al., 1989), very much like the pop- schevnikovi and the cerana clusters is ulations of , Sumatra, Sri Lanka and mainly due to characters of size. Axis 3 of S. India. A. koschevnikovi compares in cerana of the mountains a DA, which contains no characters of size size with the bees (length of vertical bars in Fig. 3), clearly in- of Yunnan (25°) and from Honan (near To- creases the distance A. koschevnikovi - kyo, 36°, Table I). A similar gradation is found in A. mellifera on A. cerana. Size in honey bees, however, is (Ruttner, 1986); highly correlated to geographic latitude this scale, A. koschevnikovi compares in size with the bees from Cameroon within the range of a species. Therefore it (4 °S) and Lamu of As- seems appropriate to introduce an ad- (coast Kenya, 2°N). ditional parameter into the taxonomic anal- suming the same scale of gradation, an A. koschevnikovi of the zone ysis : geographic latitude. All the Apis spe- temperate would be of the size of A. mellifera cies investigated so far (mellifera, cerana, (Table florea; Ruttner, 1988) clearly follow Berg- I). This, however, is a completely different mann’s rule. A correlation coefficient of scale of size variation than in A. cerana, 0.7-0.9 was found for most characters of thus furnishing another factor to distinguish size and geographic latitude in A. mellifera this species. (Ruttner, 1985). Therefore, a taxonomic The morphometric characters aside, distortion will result if equatorial popula- various others can be listed to show the tions are compared with those of higher close relation of A. koschevnikovi to A. ce- rana : tomentum on the last tergite; &dquo;indica Huttinger and Dr. R. Keller assisted in comput- vein&dquo; of the hind wing; shape of endophal- ing the graphs. lus; and pore in the capping of drone cells. It cannot be overlooked, however, that there are characters which indicate a cer- Résumé — La de l’abeille tain affinity to A. dorsata which is frequent- position rouge, Apis koschevnikovi (Buttel- ly considered as an ancestral type in rela- Reepen 1906) au sein du genre Apis. tion to the cavity-nesting species: size of a existe une science de l’abeille, honeybee in S. Asia; slender abdomen; Depuis qu’il des centaines ont été décrites high cubital index; smokey tinge of wings. d’espèces mais seules 4 ont acquis une va- Methods of molecular pro- espèces biology might leur durable. Aussi cela fit-il du bruit lors- vide further evidence of the phylogenetic qu’en 1988 Koeniger et al. et Tingek et al. among the honey bees of S. relationship montrèrent dans deux courtes notes Asia. qu’il existait à Bornéo une autre espèce nidi- As noted A. koschevnikovi is not above, fiant dans les cavités, remarquable par sa restricted to Borneo. A in the sample couleur rouge cuivre et identifiée comme Oberursel data collected 10 bank, years espèce propre en raison de la forme de ago near Solok in Sumatra and listed as l’endophallus et d’un décalage dans la pé- an isolated, &dquo;aberrant&dquo; type, clearly be- riode de vol des mâles empêchant le croi- longs to this species. The true area of dis- sement : Apis vechti (Maa, 1953). Mais, tribution of the species has yet to be inves- entretemps des recherches ont montré A certain conclusion can be made tigated. que la même abeille avait vraisemblable- about the of the The Indone- age species. ment déjà été décrite en 1906 par l’ento- sian Islands were from the con- separated mologiste berlinois H. von Buttel-Reepen the sea level rise in the tinent only during sous le nom d’A. indica koschevnikovi, post-pleistocenic period. The island popu- d’après des exemplaires de musée dont lation of A. c. indica, therefore, has not l’origine était marquée parfois &dquo;Cameroun&dquo; even reached the status of a subspecies parfois «Bornéo&dquo;. Le préfixe «indica» (au- The of the two (Ruttner, 1988). separation jourd’hui synonyme de cerana) devait indi- A. koschevnikovi and A. cerana species quer la ressemblance avec l’abeille orien- occurred in a much earlier certainly period. tale. Puisqu’aujourd’hui on sait précisément qu’A. cerana est absente d’Afrique, il faut admettre une erreur dans l’étiquetage, dans le cas où les abeilles présentent réellement des caractéristiques ACKNOWLEDGMENTS cerana. Grâce à l’amicale collaboration du Dr Frank Koch, du Museum d’Histoire Na- turelle de l’Université Humboldt à Berlin, We wish to thank Dr Frank Koch, Museum fur on a la sur des exem- Naturkunde der Humboldt-UniversitAt Berlin, Dr pu apporter preuve Christopher O’Toole, The University Museum plaires conservés intégralement, qu’ils ne Oxford, and Professor Charles Michener, Snow provenaient assurément pas d’Afrique Entomological Museum, Lawrence, Kansas, maid du sud-est asiatique. Selon la règle USA; for their kind and helpful cooperation in de priorité leur nom d’origine A. koschevni- the process of clarifying the confusing nomen- kovi doit être conservé. clature of the &dquo;Red Bee&dquo;. Agnes Mohr, Oberur- sel, was as reliable as ever in taking all the Pour l’analyse morphologique de la nou- measurements and adapting the figures. E. velle abeille, on disposait de 5 échantillons de 20 abeilles chacun récoltés dans le Zusammenfassung — Die Stellung der Nord de Bornéo (Sabah). Un autre échan- Roten Biene, Apis koschevnikovi (But- tillon de Sumatra, déjà présent dans la col- tel-Reepen 1906) in der Gattung Apis. lection d’Oberursel, a pu être déterminé Seit es eine Wissenschaft von der Honig- plus tard comme &dquo;Abeille rouge&dquo;. Les biene gibt, wurden hunderte von Bienen- échantillons de comparaison provenaient arten beschrieben, aber nur vier Arten hat- de la même collection. ten bisher dauernde Gültigkeit. Darum er- Pour déterminer la place d’A. koschevni- regte es umso größeres Aufsehen, als kovi au sein du genre Apis, on a d’abord 1988 Koeniger u.Mitarb. und Tingek u. Mi- fait une analyse discriminante (limitée aux tarb. in zwei kurzen Arbeiten zeigen konn- caractéristiques de l’aile) sur toutes les es- ten, daß es in Borneo eine weitere, pèces connues d’Apis, y compris A. arm- höhlenbrütende Art gibt, auffällig durch brusteri vieille de 12 millions d’années (Fig. ihre kupferrote Körperfarbe und als eigene 1 Les exemplaires de l’Abeille rouge se Spezies ausgewiesen durch eine andere trouvent dans un groupe individualisé, Form des männlichen Begattungs- étroitement appuyé contre A. cerana mais schlauches sowie durch eine Verschie- très éloigné d’A. mellifica. Si l’on calcule bung der Drohnenflugzeit als Kreuzungs- les ellipses de confiance à 95%, celles d’A. hemnis: Apis vechti (Maa, 1953). Inzwi- koschevnikovi et d’A. cerana se recouvrent schen angestellte Nachforschungen haben fortement (Fig. 2). aber ergeben, daß vermutlich dieselbe Si l’on analyse seules les 3 espèces Bienenart schon 1906 von dem Berliner d’Apis qui nidifient dans des cavités, les Entomologen H. v. Buttel-Reepen an Hand échantillons d’A. koschevnikovi sont situés von Museumsexemplaren, deren Herkunft entre les formes grandes (A. c. cerana) et z. T. als &dquo;Kamerun&dquo;, z.T. als &dquo;Borneo&dquo; an- les formes petites (A. c. indica) d’A. cera- gegeben war, unter dem Namen A. indica- na, bien loin des races tropicales d’A. met- koschevnikovi beschrieben worden war. lifica (Fig. 3). Les ellipses de confiance à Der Zuname &dquo;indica&dquo; (heute gleichbedeu- 95% ne montrent aucun recouvrement tend mit A. cerana) sollte auf die (Fig. 4). Si finalement on teste seule Ähnlichkeit mit der Östlichen Honigbiene l’Abeille rouge contre A. cerana avec hinweisen. Da man heute genau weiß, daß toutes ses races, une délimitation très in Afrika A. cerana nicht vorkommt, ist ein nette apparaît (Fig. 5). Irrtum in der Etikettierung anzunehmen, A. koschevnikovi est nettement plus falls diese Bienen tatsächlich cerana- Merkmale aufweisen. Dank der freundli- grande que la population locale d’A. c. ce- chen Mitarbeit von Dr. Frank Natur- rana; sa taille correspond à peu près à Koch, celle des races d’A. mellifica de la même kundliches Museum der Humboldt- latitude (Tableau 1). Puisque chez toutes Universität Berlin, konnte dieser Nachweis les espèces d’abeilles la taille augmente an den vollständig erhaltenen Exemplaren du sud au nord, on peut admettre qu’A. geführt werden, die somit sicherlich nicht koschevnikovi montrerait la même ampleur aus Afrika, sondern aus SO Asien stam- de variation de la taille qu’A. mellifica si men und nach der Prioritätsregel ihren elle avait réussi à pénétrer dans la zone ursprünglichen Namen A. koschevnikovi tempérée. behalten müssen. Zur morphometrischen Analyse der REFERENCES neuen Art standen 5 in Nordborneo (Sa- bah) gesammelte Proben zu je 15 Bienen Buttel-Reepen H. (1906) Apistica. Beitrage zur zur Eine weitere Probe aus Su- Verfügung. Systematik, Biologie, sowie zur geschichtlichen matra, die schon in der Sammlung Oberur- und geographischen Verbreitung der Honig- sel vorhanden war, konnte später eben- biene (Apis mellifica L.), ihrer VarietAten und der falls als &dquo;Rote Biene&dquo; bestimmt werden. iibrigen Apis-Arten. Veroff. Zool. Museum Berlin Die Vergleichsproben stammten ebenfalls 117-201 aus dieser Sammlung. Cornuet J.M. (1982) Representation graphique Um die Stellung von A. koschevnikovi de populations multinormales par des ellipses innerhalb der Gattung Apis zu bestimmen, de confiance. Apidologie 13, 15-20 wurde zunächst eine Diskriminanz- Koeniger N., Koeniger G., Tingek S., Mardan M. Analyse (beschränkt auf Flügelmerkmale- & Rinderer T.E. (1988) Reproductive isolation mit sämtlichen bekannten Apis-Arten, also by different time of drone flight between Apis ce- mit Einschluß der 12 Millionen Jahre alten rana Fabricius 1793 and Apis vechti Maa 1953. A. armbrusteri aus dem Randecker Mar Apidologie 19, 103-106 durchgeführt (Abb. 1 Die Exemplare der Maa T.C. (1953) An inquiry into the systematics Roten Biene sind in einer geschlossenen of the tribus Apidini or honeybees (Hymenopte- Treubia 525-640 Gruppe zu finden, eng angelehnt an A. ce- ra). 21, rana, aber weit entfernt von A. mellifera. Rinderer T.E., Koeniger N., Tingek S., Mardan Berechnet man den Vertrauensbereich der M. & Koeniger G. (1989) A morphological com- of the bees of Bor- statistischen Abgrenzung bei 95%, so er- parison cavity dwelling honey gibt sich eine starke Überlappung der kon- neo, Apis koschevnikovi (Buttel-Reepen, 1906) and cerana von A. koschevnikovi und A. Apis (Fabricius 1793). Apidologie fidenzellipsen 20, 405-4111 cerana (Abb. 2). Ruttner F. Graded Werden die drei höhlenbrütenden (1985) geographic variability Apis- in honeybees and environment. Pszcz. Zesz. Arten für sich analysiert, so liegen die Por- Nauk29, 81-92 ben von A. koschevnikovi zwischen den Ruttner F. (1988) Biogeography and Taxonomy c. und den kleinen For- großen (A. cerana) of Honeybees. Springer Verlag, Heidelberg men (A. c. indica) von A. cerana, weit ent- Ruttner Tassencourt L. & Louveaux J. fernt von Rassen von A. mellif- F., (1978) tropischen Biometrical-statistical of the era Die von analysis geographic (Abb. 3). Konfidenzellipsen variability of Apis mellifera. Apidologie 9, 363- 95% zeigen keine Überlappung (Abb. 4). 381 Wird schließlich die Rote Biene gegenüber Ruttner F., Wilson R., G., Vorwohl G. & A. cerana mit all ihren Rassen allein Snelling getes- Kauhausen D. (1986) The evolution of the hon- so sich eine sehr klare tet, ergibt Abgren- eybee wing venation. Apidologie 17, 349 zung (Abb. 5). Smith F. (1858) Catalogue of the Hymenopter- A. koschevnikovi ist ganz wesentlich ous collected at Sarawak, Borneo, größer als die lokale Population von A. ce- Mount Ophir, Malacca, and at Singapore, by rana indica; sie entspricht in ihrer Größe A.R. Wallace. J. Proc. Linn. Soc. London Zool. etwa den Rassen von A. mellifera aus der- 2, 42-130 selben geographischen Breite (Tab. I). Da Smith F. (1865) On the species and varieties of bei allen Bienenarten die Formen von the honey-bee belonging to the genus Apis. Süden nach Norden zu größer werden, Ann. Mag. nat. Hist. London 15, 372-380 kann man annehmen, daß A. koschevniko- Tingek S., Mardan M., Rinderer T.E., Koeniger vi dieselbe Größen-Variations breite wie A. N. & Koeniger G. (1988) Rediscovery of Apis mellifera hätte, falls es ihr gelungen wäre, vechti (Maa: 1953): the Saban honey bee. Api- in die gemäßigte Zone vorzudringen. dologie 19, 97-102