© INRA/DIB-AGIB/EDP Sciences, 2009

1. INTRODUCTION

In recent years the dwarf honeybee Apis florea Fabricius has been steadily expanding westwards, both naturally and inadvertenetly via global transportation (Mogga and Ruttner, 1988; Hepburn et al., 2005). It is now well established in Iraq, Oman and Yemen (Wongsiri et al., 1996); and has recently been detected in Sudan, where it was first detected in 1985 (Mogga and Ruttner, 1988) and in central Saudi Arabia (Hepburn et al., 2005). Throughout this expansion along the Arabian peninsula and into Africa it has proven to be a highly successful colonizer, well adapted to hot arid conditions of both urban and rural landscapes and seemingly unaffected by competition from any local A. mellifera (El-Shafie et al., 2002). Most recently, A. florea has become established in the area around Aqaba, Jordan (Haddad et al., 2008). This region is about 1500 km apart from its next occurrences in Sudan, and about 2000 km apart from its next natural occurrences in Iran and Oman. We here report the results of morphometric analyses on the Jordanian and Sudanese samples in combination with others from the western half of the A. florea distribution to identify the probable source of origin of the founder population in Jordan.

The expansion of the range of A. florea is of more than academic interest given the expansions of the ranges of A. cerana Fabricius and A. mellifera L. and their possible effects on endemic biodiversity. In the last century the invasion of the Americas by the African A. m. scutellata Lepeletier has displaced the previously invasive European races of A. mellifera with yet uncharted consequences for local apifauna and other insect pollinators (Winston et al., 1981; Soares and De Jong, 1992). Similarly, since its introduction into Papua New Guinea, A. cerana has spread throughout this great island and is poised to enter Australia over the stepping stone small islands of the Torres Strait (D. Anderson, unpubl. data). Recommended policy of the Australian Honeybee Industry Council (2008, http://www.quarantinebiosecurityreview.gov.au/submissions_received) is to contain the introduction of another invasive honeybee species because of possible effects on biodiversity as well as commercial beekeeping on that continent. Against these concerns, it has been reported that while A. florea is indeed rapidly spreading through Sudan, that there are no real competitive interactions between the native A. mellifera and the former (El-Shafie et al., 2002). Possible effects on the local apifauna and other insect pollinators remain unexplored.

2. METHODS AND MATERIALS

2.1. Honeybees

The worker honeybees used for the analyses of A. florea in this study derive from: (1) new material collected from Jordan (9 colonies) and Sudan (8 colonies); (2) raw data from Southern India (1 colony), Sri Lanka (6 colonies), Iran (19 colonies), Pakistan (3 colonies), Saudi Arabia (1 colony) and Oman (3 colonies) from the Institut for Bienenkunde database at Oberursel. Morphometric analyses were performed on 870 individual worker bees from 50 colonies representing 28 localities (Fig. 1 and additional online material for the geographical co-ordinates).

Figure 1 Geographical localities at which the Apis florea used in the study were collected. Names of the localities and their co-ordinates are given in additional online material.

2.2. Measurements

Morphological characters of worker bees were measured using the Ruttner (1988) parameters as follows: length of femur (5), tibia (6), metatarus (7), metatarsus width (8), length of tergite 3 (9), tergite 4 (10), forewing length (17), forewing width (18), cubital a (19), cubital b (20). Angles of venation were not available for most colonies from Iran and size related characters were not available for the colonies from Pakistan and Sri Lanka.

2.3. Data analysis

Colony sample means and standard deviations were computed for each morphometric character from 15-20 bees per colony. Multivariate statistical analysis of the data included principal components, discriminant and k-means cluster analyses (Johnson and Wichern, 2002). All tests were performed using Statistica (StatSoft, 2007).

3. RESULTS

Principal components analyses were carried out using the colony means of ten morphometric characters of worker honeybees ((5), (6), (7), (8), (9), (10), (17), (18), CI = (19)/(20)). Two principal components with eigenvalues greater than one were isolated. PC 1: size-related characters (5) to (10), and (17) and (18) with component loadings between 0.86 and 0.96 accounted for 69.7% of the variation; PC 2: cubital index (CI) and size (8) with component loadings 0.90 and 0.53, respectively accounted for 14.1% of the variation. The two principal components accounted for a total of 83.8% of the variation in the data. The PC plot using the first and second PC scores, revealed two morphoclusters with a group of smaller bees from S. India and Sri Lanka forming one cluster, and an isolation of larger bees from Iran, bees from Sudan, Oman, Jordan and Pakistan forming the second cluster.

Canonical 1 and 2 scores plot from a discriminant analysis using the colony means of the same ten morphometric characters and the six countries with n > 1 as the groups showed the colonies from Jordan and Sudan within the same 90% confidence ellipse as those from Pakistan and Oman (Fig. 2). The results of the squared Mahalanobis distances between the centroids of the groups showed that the colonies from Jordan were likely introduced via Oman (d² = 4.6, Tab. I).

Figure 2 Discriminant analysis plot using the colony means of ten morphometric characters. Letters indicate colonies from countries: SL = colonies from Sri Lanka (cluster 1), J = colonies from Jordan, O = colonies from Oman, P = colonies from Pakistan, S = colonies from Sudan (cluster 2), I = colonies from Iran (cluster 3). Confidence ellipses are drawn at the 90% level.

A cluster analysis using the complete linkage procedure was carried out on colony mean character values combined for the 8 countries. The results showed a dendrogram of two main clusters. Cluster 1 first linked colonies from Pakistan and Sudan, then Oman and Jordan and then Saudi Arabia and finally Iran; cluster 2 linked colonies from S. India and Sri Lanka (Fig. 3).

Figure 3 Hierarchical clustering dendrogram for A. florea in Jordan and adjacent countries, derived from complete linkage clustering procedure on ten morphometric characters averaged for countries.

4. DISCUSSION

The morphometric analyses clearly indicate that the A. florea in Jordan is derived from Oman. Though no sampling locations further to the east than Sri Lanka were included, the firm placement of the Aqaba samples within the Iran-Pakistan and Arabic Peninsula samples excludes any affiliation with A. florea from far eastern origins. The analyses also reconfirm that Pakistan was the origin of A. florea in Sudan (Mogga and Ruttner, 1988). It is almost certain that the A. florea in Jordan arrived through human transportation because the region is surrounded by vast expanses of barren wasteland that precludes gradual expansion of the natural geographic range. It is very probable that the route of entry was by sea into the port of Aqaba in the northeastern hook of the Red Sea. And, indeed, there is a very large shipping trade along the Red Sea and Arabian Sea between Oman and Jordan. The A. florea in Sudan were first discovered in Khartoum but have since spread along the Nile River (Mogga, unpubl. data). It seems highly probable that the Jordan and Sudanese populations will eventually converge in the Sinai Peninsula.

Recording the occurrence of a biogeographical event such as the introduction of A. florea to Jordan may appear insignificant. However, in an ecological and conservation context, such events could have considerable effects. There is a nascent but growing corpus of literature documenting the effects of foreign introductions, many of which achieve major pest status (Samways, 1994; Kearns et al., 1998; Kato and Kawaki, 2004; Abe, 2006). Indeed, the history of applied entomology can be viewed as a chronicle of the behaviour of introduced species which have become invasive (Howard, 1930; Pedigo and Rice, 2008). We suggest that, short of eradication of such invasive honeybees, their careful scrutiny in a new environment is certainly warranted.

Acknowledgments

Many thanks to A. Bataene, E. Algareb, H. Amawe and G. Obedat for assistance in locating and sampling the colonies.

References

• Abe T. (2006) Threatened pollination systems in native flora of the Ogasawara (Bonin) Islands, Ann. Bot. 98, 317–334 [CrossRef] [PubMed].
• El-Shafie H.A.F., Mogga J.B.B., Basedow T.H. (2002) Studies on the possible competition for pollen between the honey bee, Apis mellifera sudanensis, and the imported dwarf honey bee Apis florea (Hym., Apidae) in North-Khartoum (Sudan), J. Appl. Entomol. 126, 557–562 [CrossRef].
• Haddad N., de Miranda J.R., Bataena A. (2008) Discovery of Apis florea in Aqaba, Jordan, J. Apic. Res. 47, 173–174.
• Hepburn H.R., Radloff S.E., Otis G.W., Fuchs S., Verma L.R., Tan K., Chaiyawong T., Tahmasebi G., Wongsiri S. (2005) Apis florea: morphometrics, classification and biogeography, Apidologie 36, 359–376 [CrossRef] [EDP Sciences].
• Howard L.O. (1930) A history of applied entomology, Smtihsonian Institution, Washington.
• Johnson R.A., Wichern D.W. (2002) Applied Multivariate Statistical Analysis, Fifth Edition, Prentice Hall, Upper Saddle River, New Jersey.
• Kato M., Kawakita A. (2004) Plant-pollinator interactions in New Caledonia influenced by introduced honey bees, Am. J. Bot. 91, 1814–1827 [CrossRef].
• Kearns C.A., Inouye D.W., Waser N.M. (1998) Endangered mutualisms: The conservation of plant-pollinator interactions, Annu. Rev. Ecol. Syst. 29, 83–112 [CrossRef].
• Mogga G.B., Ruttner F. (1988) Apis florea in Africa; source of the founder population, Bee World 69, 100–103.
• Pedigo L.P., Rice M. (2008) Entomology and pest management, Pearson, London.
• Ruttner F. (1988) Biogeography and taxonomy of honeybees, Springer, Berlin.
• Samways M.J. (1994) Insect conservation biology, Chapman & Hall, London.
• Soares A.E.E., De Jong D. (1992) Pesquisas com Abelhas no Brasil, Revista. Brasileira de Genetica., Ribeirao Preto SP, 680 p.
• StatSoft, Inc. (2007) STATISTICA, version 8.1, www.statsoft.com (accessed on 2 March 2009).
• Winston M.L., Dropkin J.A., Taylor O.R. (1981) Demography and life history characteristics of two honey bee races (Apis mellifera), Oecologia 48, 407–413 [CrossRef].
• Wongsiri S., Lekprayoon C., Thapa R., Thirakupt K., Rinderer T.E., Sylvester H.A., Oldroyd B.P., Booncham U. (1996) Comparative biology of Apis andreniformis and Apis florea in Thailand, Bee World 77, 23–35.

Online material

Download PDF file.

All Tables

All Figures

	Figure 1 Geographical localities at which the Apis florea used in the study were collected. Names of the localities and their co-ordinates are given in additional online material.
In the text

	Figure 2 Discriminant analysis plot using the colony means of ten morphometric characters. Letters indicate colonies from countries: SL = colonies from Sri Lanka (cluster 1), J = colonies from Jordan, O = colonies from Oman, P = colonies from Pakistan, S = colonies from Sudan (cluster 2), I = colonies from Iran (cluster 3). Confidence ellipses are drawn at the 90% level.
In the text

	Figure 3 Hierarchical clustering dendrogram for A. florea in Jordan and adjacent countries, derived from complete linkage clustering procedure on ten morphometric characters averaged for countries.
In the text