Bee Conservation in Sub-Saharan Africa and Madagascar: Diversity, Status and Threats Connal D

Bee conservation in Sub-Saharan Africa and Madagascar: diversity, status and threats Connal D. Eardley, Mary Gikungu, Michael P. Schwarz

To cite this version:

Connal D. Eardley, Mary Gikungu, Michael P. Schwarz. Bee conservation in Sub-Saharan Africa and Madagascar: diversity, status and threats. Apidologie, Springer Verlag, 2009, 40 (3), 10.1051/apido/2009016. hal-00892023

HAL Id: hal-00892023 https://hal.archives-ouvertes.fr/hal-00892023 Submitted on 1 Jan 2009

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. Apidologie 40 (2009) 355–366 Available online at: c INRA/DIB-AGIB/EDP Sciences, 2009 www.apidologie.org DOI: 10.1051/apido/2009016 Review article

Bee conservation in Sub-Saharan Africa and Madagascar: diversity, status and threats*

Connal D. Eardley1,MaryGikungu2, Michael P. Schwarz3

1 Agricultural Research Council, Private Bag X134, Queenswood, 0121, Pretoria, South Africa 2 Zoology Department, National Museums of Kenya, PO Box 40658-00100, Nairobi, Kenya 3 School of Biological Sciences, Flinders University, GPO Box 2100, Adelaide, South Australia 5001, Australia

Received 14 October 2008 – Revised 2 February 2009 – Accepted 4 February 2009

Abstract – Sub-Saharan Africa and Madagascar contain a wealth of bee diversity, with particularly high levels of endemicity in Madagascar. Although Africa contains seven biodiversity hotspots, the bee fauna appears rather moderate given the size of the continent. This could be due to various factors, an important one being the dearth of bee taxonomists working in Africa and diﬃculties in carrying out research in many regions. Anecdotal observations suggest a very large number of undescribed bee species. A number of se- rious threats to this diversity exist, especially habitat destruction and degradation. Bee diversity in these regions is likely to be important for both agriculture and indigenous ecosystems, but is under-appreciated. Reliance on conserved areas such as National Parks will not be suﬃcient to preserve bee diversity in Africa and Madagascar; changes to land use practices and development of industries that facilitate conservation, such as ecotourism, will be essential. There is also a strong need to build regional expertise and infrastructure that can be used for documenting bee diversity, identifying the most urgent conservation issues, and implementing conservation strategies. Support from developed countries and international funding agencies is needed for this. bees / conservation / biodiversity / Africa / Madagascar / Apoidea

1. INTRODUCTION ing materials, the beetle larvae that San people used for poison arrows (Koch, 1958;Shaw Trying to understand the issues surrounding et al., 1963) and harmful species like agricul- bee conservation in sub-Saharan Africa and tural pests. They have names for all of these Madagascar is complex. It requires consider- species. Food and medicinal uses for honey ation of both the history of bee research in from stingless bees are known (Macharia, un- these regions, as well as the prevailing socio- publ. data), but today these are better known to economic and cultural circumstances. These rural communities than urban people. are vastly different from North America and Europe, and the strategies that are needed to While honey bees and sometimes stingless conserve bee diversity are similarly very dif- bees are widely regarded for their pollina- ferent (cf. Byrne and Fitzpatrick, 2009). tion services, the benefits from other bees are African and Malagasy people appreciate the largely overlooked by farmers – to what ex- aspects of biodiversity that tangibly affect their tent they were recognised in the past is not daily lives. These include beneficial species, known. Increased crop production has been such as those used for medicine, food, build- mainly based on the use of agronomic inputs, such as quality seeds, fertilizers and pesticides, Corresponding author: M.P. Schwarz, with little regard to non-Apis pollination as an Michael.Schwarz@flinders.edu.au agricultural eco-service that may require pro- * Manuscript editor: Tomas Murray tection.

Article published by EDP Sciences 356 C.D. Eardley et al.

Figure 1. Map of Sub-Saharan Africa showing broad distribution of vegetation types. Adapted from the NASA Scientiﬁc Visualization Stu- dio maps (http://svs.gsfc.nasa.gov/ vis/) and the University of Chicago Fathom Archive African map series (http://fathom.lib.uchicago.edu/).

Bee research in sub-Saharan Africa has in- deals with the protection of Kakamega Forest cluded extensive research on the honey bee, in Kenya. Apis mellifera L., especially in South Africa Although Africa and Madagascar are ge- (Hepburn and Guye, 1993; Johannsmeier, ographical neighbours, they are in many 2001), limited work on stingless bees, haphaz- respects very different. They are therefore ard descriptions of new species, revisional tax- mostly treated separately in the discussions onomic studies for two-thirds of the known that follow. genera, a host of distributional records, and a relatively small number of studies on ecology and social biology. Eardley and Urban 2. BEE DIVERSITY (in preparation) provide complete references to the non-Apis studies in a species catalogue. 2.1. Sub-Saharan Africa Research on Malagasy bees has been simi- lar, except that all previous studies have been Sub-Saharan Africa is an enormous and di- brought together with a full revision of the bee verse continent with a host of different ecosys- fauna by Pauly et al. (2001). tems, from rain forest to desert. A broad veg- Prospects for bee research in sub-Saharan etation map for regions south of the Sahara is Africa are improving, and there is a growing given in Figure 1. A large proportion of the awareness of the need to conserve pollinator continent is savannah, with many different bee diversity. This is largely driven by an interest pollinated plants, and with biotypes ranging in agricultural pollination management. How- from dry to relatively wet. The continent has ever, the potential importance of all bees, as seven biodiversity hotspots, as identified by part of an ecosystem approach to conserva- Conservation International (CI). (http://www. tion, is recognized in a number of projects. biodiversityhotspots.org), making it second to One such project is BIOTA East Africa, which Asia and the Pacific Islands in biologically rich Bee conservation in Africa and Madagascar 357 areas. The highest temperate plant biodiversity in the world occurs in the winter rainfall areas in the south-western region of Africa (Koekemoor, unpubl. data). All this should suggest a rich bee fauna. The described bee fauna in sub-Saharan Africa can, however, at best be described as moderately diverse. Six of the seven bee families recognized by Michener (2007) occur in Sub-Saharan Africa, with the Stenotriti- dae being confined to Australia. About 21% of the World’s bee genera occur in Sub- Saharan Africa (102 from a total of about 476; Michener, 2007), and about half of these are either cosmopolitan or Old World en- demics. At a generic level, this suggests rea- sonably high diversity. However, at a species level diversity is less rich. The 2600 described Sub-Saharan species comprise only about 13% of the global fauna of around 19400 species (http://www.itis.gov./ and Eardley, unpubl. data). Some highly speciose genera, like Perdita and Centris, do not occur in Africa, whilst Andrena, which is diverse in the Ho- larctic (about 1500 species), has only nine African species. The most speciose genera in Africa are Lasioglossum (about 260 species, Figure 2. Map of Madagascar showing broad vege- no revision available) and Megachile (about tation types (excluding marshlands). Adapted from 322 species, Pasteels, 1965), and these are the Food and Agriculture Organization (United fairly small proportions of their global species Nations) (http://www.fao.org/ag/AGP/agpc/doc/ numbers. Counprof/Madagascar/madagascareng.htm)and It is possible that low species diversity in the University of Texas Perry-Castañeda Map Africa is more apparent than real. Schwarz Collection (www.lib.utexas.edu/maps/madagascar. and Bull (unpublished) found that in extensive html). nest collections of allodapines in South Africa, about one third of species encountered did not fit current species descriptions, and less- 2.2. Madagascar intensive allodapine collections from Uganda indicate a great many undescribed species and Madagascar is the World’s 4th largest is- possibly new genera (McLeish and Schwarz, land. It exhibits a much wider range of ecosys- unpubl. data). For allodapines at least, this tems and biodiversity than most other large suggests a wealth of unrecorded diversity, de- islands. This is probably due to its physi- spite the Herculean efforts of Michener (1975) cal geography, the times since it rifted from in revising the African allodapines. The prob- other major land masses, and its distance from lem appears to be one of a mismatch be- sources for cross-ocean dispersal (Yoder and tween the sheer size of sub-Saharan Africa Nowak, 2006). The eastern mountain range and the number of African bee taxonomists. also leads to very different climatic regimes Even with the involvement of taxonomists across the island. There is lowland rainforest from other regions, political and infrastruc- in the east, montane and dry deciduous forest tural circumstances make biodiversity research in the west and high elevation thicket in the very difficult in many important regions. central highlands (Fig. 2). The south-west is 358 C.D. Eardley et al. arid with spiny forests and scrub. All regions represent dispersal via vegetation rafts aris- have suffered from prolonged slash and burn ing from tsunamis or extreme floods in east- agriculture, leading to extensive grassland and ern Africa, since several large rivers flow into areas dominated by exotic weeds such as Eu- the Mozambique Channel separating Africa calyptus and Grevillea. from Madagascar. On the other hand, another Madagascar has played an important role in wood nesting bee, Lithurgus pullatus, occurs our understanding of historical biogeography in Madagascar, East Africa, India, Mauritius, (Yoder and Novak, 2006). However, for many the Seychelles and the Maldives (Pauly et al., groups of organisms it is still not clear whether 2001), and it is unlikely this could be ex- affinities with African taxa represent vicariant plained by vegetation rafting. The possible origins, dating back to the rifting of Madagas- role of anthropogenic dispersal for some bees car+India from Africa, or whether this is due in Madagascar is something that needs study. to dispersal across the Mozambique Channel. The only ceratinine recorded from Mauritius The combined land mass of Madagascar and is identical to Ceratina dentipes from Malaysia India began rifting from Africa about 165– and is likely to represent a dispersal via ocean 155 my before present (BP) and Madagascar trading routes (Rehan and Schwarz, unpubl. reached its current position relative to Africa data). about 130–118 my BP. It is generally thought that bees first evolved in the early-mid Creta- ceous, perhaps 120 my BP (Engel, 2001), so 3. BEE DISTRIBUTION a vicariant origin for any endemic Malagasy AND ENDEMICITY bee clades is unlikely. It has been hypothesized that a land bridge, or island ‘stepping stones’, 3.1. Sub-Saharan Africa connected Madagascar to Africa from the mid- Eocene to the late Miocene (McCall, 1997), Sub-Saharan Africa’s bee diversity is great- though such a dispersal corridor is contentious est in the winter rainfall area of South Africa. (Rogers et al., 2003). This area is species-rich for all families ex- The likelihood that Madagascar contains re- cept Andrenidae, and this fits with global pat- lictual elements from the earliest radiations of terns showing high bee diversity in Mediter- the bees seems low. There are no Malagasy ranean climates (Michener, 2007). However, a species in the most basal bee family Dasypo- number of taxa are most diverse in the trop- daidae (Danforth et al., 2006) and in the next- ics, with decreasing diversity at higher lati- most basal clade, Meganomiidae + Mellitidae tudes. The large carpenter bees (Xylocopa)are s. s., there are only two Malagasy species, both one example of this phenomenon, with many meganomiids (Pauly et al., 2001). The only species are confined to the tropics, and other stingless bee genus in Madagascar, Liotrig- species’ ranges extending southwards to vary- ona, is not basally situated in the Meliponini ing extents. Some extend down the moun- (Rasmussen and Cameron, 2007), and all the tain ranges, like Xylocopa somalica Magretti, Malagasy allodapine and ceratinine lineages and others extend to different degrees along originated from dispersals from Africa, begin- the coast, with the range of Xylocopa nigrita ning about 30 Mya and continuing up until a (Fabricius) extending to northern Kwazulu- couple of million years ago (Fuller et al., 2005; Natal and with Xylocopa flavorufa (DeGeer) Schwarz et al., 2006; Smith et al. 2007; Rehan reaching the southern Cape. The South West- and Schwarz, unpubl. data). ern Cape has several unique, endemic species, Some elements of the Malagasy bee fauna like Xylocopa albifrons Lepeletier and Xylo- show evidence for high levels of vagility copa capitata Smith. among the Mascarene islands and the Indian A bee few genera, like Dactylurina Cock- Ocean Rim. Three of the six Malagasy Xylo- erell, are exclusively tropical, but several are copa species also occur in Africa, and a fourth endemic to southern Africa. Some of these are occurs in India (Pauly et al., 2001). It is pos- desert bees, like Mermiglossa Friese, which is sible that these African-Malagasy connections confined by the tropics and savannah to the Bee conservation in Africa and Madagascar 359 north and the east of its range. Southern Africa are major impediments to assessing even basic has a few endemic genera that occur in both elements of bee diversity in much of tropical the dry, winter rainfall areas in the west and Africa and the Horn of Africa. the wetter, summer rainfall areas in the east, like Rediviva Friese. A few genera are diverse in the Palaearctic and recorded in southern 3.2. Madagascar Africa from only one or a few rare species, such as Aglaoapis Cameron. Fidelia Friese, Madagascar has a very high level of bee en- on the other hand, has its centre of diversity demicity compared to the rest of the world, in southern Africa with a relic species in Mo- with about 90% of species being endemic and rocco. Neither of these genera have any known with ten endemic genera (out of 86 genera species in the tropical regions of Africa. Meli- on the island) and another five endemic sub- turgula Friese has its maximum diversities in genera. There are about 240 described species southern Africa and the Sahara Desert, yet one from six families. Of the 18 bee tribes recorded species, Meliturgula scriptifrons (Walker), oc- from Madagascar (Michener, 2007), all are curs in East Africa. The Anthidiini and Osmi- found in Africa, so the Malagasy bee en- ini, which are largely cosmopolitan, have their demism does not extend to taxa higher than centres of diversity in the south-western region genera. Given that India rifted from Mada- of Africa. gascar well after Madagascar separated from Very little is known about the bee fauna of Africa, it is interesting that there are no genera north-east Africa, but it is likely to be an inter- or tribes in Madagascar that occur in Asia but esting region for endemism. For example, six not Africa. of the ten species of Nomada Scopoli occur The Malagasy bee fauna is dominated by in Ethiopia, with only three species in South the Halictidae (123 species) and the Apidae Africa. The region between tropical Africa and (82 species) and nine of the ten endemic gen- the Sahara Desert is narrow and does not ap- era belong to these two families (the other pear to have the diverse ecosystems that occur endemic genus is a megachilid). Three fam- south of the tropics. However, this area is very ilies have very minimal representation: there poorly sampled and political situations in the are only three colletid species (all in the Hy- Horn of Africa mean this situation is unlikely laeinae), one andrenid (a Meliturgula species to be rectified soon. closely related to M. scriptifrons [Pauly et al., Both at generic and species levels there 2001]) and two meganomiids (Pauly et al., appears to be higher diversity and endemic- 2001). ity in the CI hotspots. Rediviva and Col- The only comprehensive treatment of the letes have distinct south-western Cape (Cape Malagasy bee fauna is Pauly et al. (2001). Floristic Region and Succulant Karoo, Fig. 1) This study is remarkable for the extent of its and south-eastern Cape centres of diversity coverage, the number of new taxa described, (Whitehead and Steiner, 2001; Whitehead and the extensive provision of biogeograph- et al., 2008; Kuhlmann, 2005). East Africa ical records. More recently, the phylogenetic has several unique melittid genera (Michener, relationships of Malagasy species to elements 1981) and given the basal nature of this group from other regions have been explored for two among bees (Danforth et al., 2006)thisis tribes using molecular data. Danforth et al. likely to become a critical biodiversity re- (2008) showed that in the Halictini, the Mala- source for inferring early stages in bee evolu- gasy species of Patellapis were all nested tion. within an African clade, and that one Asian Overall, coming to terms with bee diver- species had a well-supported sister-clade rela- sity and endemicity in sub-Saharan Africa is tionship to a Malagasy group. For the allodap- problematic because the continent is largely ine bees, Fuller et al. (2005) found that there under-collected and is likely to remain so, par- were two colonisation events of Madagascar ticularly in the north east, for some time. Po- by the genus Braunsapis Michener, one about litical and infrastructural barriers to research 2.6 my BP and another one about 12.8 my BP. 360 C.D. Eardley et al.

On the basis of larval morphology and se- of niche specificity in east Africa involves quence data, Chenoweth et al. (2008)trans- species of Ctenoplectra – oil collecting bees ferred Malagasy Halterapis Michener species with highly modified hind legs characterized to a new genus, Hasinamelissa. Molecu- by long brush-like hairs on the tibia. They lar data suggest that this lineage diverged forage mainly on Mormodica foetida, an oil- from an African clade at least 20 my BP producing plant. Ctenoplectrina, which has (Chenoweth and Schwarz, 2007). Seventeen two cleptoparasitic species, are found in close species of Hasinamelissa have been described, association with Ctenoplectra (Eardley, 2003). and Chenoweth et al. (2008) indicate the ex- Unfortunately little is known about their nest- istence of another three undescribed species ing biology. These two genera are mainly based on DNA sequence data. That study was found together in areas rich with oil produc- based on limited collections centred on major ing plants such as Mormodica foetida and M. provincial towns and it is likely that the ac- balsamina (Eardley, 2003; Gikungu, 2006). tual number of extant species is much larger than currently described. In fact, the number of described allodapines in Madagascar is now 4.2. Madagascar larger than for all of Asia (Reyes, 1991), making Madagascar the third most speciose region Malagasy bee locality records (Pauly et al., for allodapines in the world, after Sub-Saharan 2001) indicate a wide range in distributional Africa and Australia. patterns, ranging from near ubiquity to very limited distributions, sometimes comprising single locality collections. More extensive col- 4. NICHE SPECIFICITY lections are needed to confirm whether the single locality records really do indicate ex- tremely limited distribution. There are few 4.1. Sub-Saharan Africa data that allow distributions to be linked with specific resource or climatic require- In Sub-Saharan Africa some bees are gener- ments, though most genera and subgenera alists, such as Xylocopa caffra, which occurs have marked regional distributions. There ap- in winter and summer rainfall areas, deserts, pear to be diversity hotspots that roughly savannah and forests, nesting in a variety of correspond to some habitat types. Particu- woods and pithy plant stems (Eardley, un- larly notable areas are the spiny forests and publ. data). X. sicheli, on the other hand, scrub in the far south west, the forests in nests only in dead Aloe inflorescences and is the Ranomafana region, the remote Morarano- confined to dry savannah and semi-desert ar- Chrome region in the north east, the coastal eas (Eardley unpubl. data). Xylocopa scioensis area south of Tamatave, and the Morondava re- Gribodo nests only in dead Phragmites spp. gion (Fig. 2). However, because of the exten- stems, and therefore only occurs near water sive degradation of native vegetation, it is not courses. Rediviva spp. (Mellitidae s. s.)col- clear whether current diversity hotspots reflect lect oil from a range of orchids as well as the original distributions, rather than remnants of genus Diascia (Scrophulariaceae) using modi- indigenous vegetation. fied front tarsi. Their front legs are of comparable length to the oil producing spurs of the flowers (Whitehead et al., 1984; Whitehead, 5. CHANGES TO ECOSYSTEMS unpubl. data). The plants visited by bees in the arid and semi-arid areas of southern Africa 5.1. Africa have been well documented by Gess and Gess (2003). Africa, as with Europe, is home to the Most African bee communities remain honey bee, Apis mellifera Linnaeus and is the poorly understood in terms of pollinator di- birthplace of modern humans (Stringer and versity and niche specificity – largely due Andrews, 2005). Although these factors im- to lack of extensive sampling. One example pact on bee diversity and abundance, they Bee conservation in Africa and Madagascar 361 should probably be considered part of the eco- arid and semi-parts of eastern Kenya and has logical ground plan for Africa. Nevertheless, reduced the population of both arboreal and technologically modern man has contributed ground nesting bees (Gikungu and Njoroge, to pollinator declines in Africa. This is evident 2007). in that bee diversity and abundance is much There have been no studies that examine greater on crops in areas surrounded by natural whether loss of bee diversity impacts on horti- vegetation than in ecosystems that have been cultural yields. However, a few studies in east widely transformed by agriculture (Eardley, Africa (Gikungu et al., 2008)haveshownthat unpublish. data). agricultural ecosystems may support higher Honey bees are an important pollinator of levels of bee diversity and abundance than the African flora and part of the pollinator forested areas. This is because disturbed habi- guilds of natural ecosystems, but they can tats in low-intensity farming systems may pro- displace other bees on flowers (Eardley, un- vide more suitable resources for many bee publ. data). Their natural abundance is a moot species. This calls for further holistic ap- point. San people paintings in the Drakens- proaches in regard to management of disturbed berg, South Africa, indicate that honey hunt- and undisturbed habitats given that bees are ing has taken place for centuries. This was both coarse and fine-grained organisms. by destructive sampling that could have reduced wild honey bee numbers. However, the honey bee is the only managed pollinator 5.2. Madagascar in Africa, and their numbers and distribution have increased with domestication and the in- Madagascar has suffered some of the worst troduction of alien plants, like Eucalyptus spp. destruction of natural ecosystems in the world. (Wiese, unpubl. data). At the same time the re- This began approximately 2000 years ago with cent movement of the Cape Honey Bee into human colonisation from southern Asia and Gauteng Province, South Africa, and the intro- Polynesia, and a major problem derives from duction of honey bee diseases, such as Varroa, extensive slash and burn agriculture. The con- have reduced honey bee numbers over the past tinuing threat from this is still evident to any- two decades. one flying over Madagascar, where the tell- The most comprehensive work on bee con- tale signs of fire and smoke forming thin servation in Africa is that of Gess and Gess lines at the edges of forests are depressingly (1993). They outline factors that affect the common. This has left most of the central abundance and diversity of bees in semi-arid hauts plateaux comprising a mixture of an- areas, but their findings are likely to be more thropogenic grasslands and eucalyptus planta- widely applicable. Transformation of the land- tions that are cropped to provide small timber scape by replacement of natural vegetation for local construction and charcoal. The result with crops and other exotics, along with re- is also evident in wide-scale erosion, evident moval of natural vegetation through urban- from red plumes extending tens of kilometres ization, are major problems. These problems into the sea from the major estuaries. are greatest in the higher rainfall regions, pri- The effects of land clearing on bee diver- marily the tropical, subtropical, and montane sity is suggested by a comparison by Pauly areas, along with the Mediterranean ecosys- et al. (2001) of bee species between coastal tems of South Africa where extensive crop- and montane forests with a coastal urban and ping has occurred. More subtle threats arise montane cultivated area. The former habitats from selective grazing, removal of bush to ex- had 49 and 69 species respectively while the pand grasslands, trampling by stock, removal latter two regions had 25 and 20 species. That of wood and charcoal production, and the use study did not provide a firm quantitative ba- of insecticides to control diseases like sleep- sis where seasonal and other proximate effects ing sickness. These problems are important were controlled for, but is nevertheless alarm- in savannah and semi-arid areas. Slash and ing. There is a strong need for further studies, burn agriculture is also a common practice in including studies that examine the effects of 362 C.D. Eardley et al. exotic weeds which, in some large areas, have of action, through the African Pollinator Initia- almost entirely replaced native vegetation. tive (API, Martins et al., 2003), for a pollinator conservation consortium. The main goal of API is to promote pollination as an essential 6. THREATS ecosystem service for sustainable livelihoods and the conservation of pollinator biological diversity (Martins et al., 2003). However, the The same general suite of threats to pollina- only strategies specifically regarding bee con- tor decline affect both Africa and Madagascar: servation that have been implemented to date namely, spread of exotic weeds, pesticide use are designed to protect agriculture and not bee and, most importantly, habitat destruction and diversity per se. One impediment to bee con- degradation. Perhaps the main driver of these servation is that while honey bees are per- threats comes from population growth and the ceived to be important pollinators, this does imperative to improve living standards, com- not seem to extend to other bee groups. In bined with economic circumstances that make a continent where government finances are the implementation of conservation strategies often very limited and socio-economic and very difficult. health challenges are very large, establishing In Madagascar there has been a massive in- a priority for bee conservation will be diffi- crease in the number of national parks since cult. Outside of protected areas like national the 1980’s, but in both Madagascar and Africa parks, making agricultural practices more bee- there is a need for smaller-scale conservation friendly will be important, but there is a need strategies that have the support of local com- for studies to show that such measures will munities. Conservation measures imposed by actually benefit communities if they are to be central governments may lack effectiveness if widely adopted. the benefits to local communities are not evident to those communities, and if the measures cannot be policed. To this extent, there need to be clear links between conservation and the 7.2. Madagascar well-being of local communities. Ecotourism and development of sustainable agriculture are For Madagascar the only comprehensive two ways in which these links can be encour- stock-take of indigenous bee was by Pauly aged, but we should not underestimate the size et al. (2001), remarkable for its scope and of the challenge. It seems unlikely that conser- highlighting of biodiversity issues. That study vation measures will be effective if they ignore could form the basis for a more focussed ex- social issues. amination of conservation risks and strategies. Regional maps of species records provided by Pauly et al. (2001) could be used to identify areas where diversity is high, and this synthe- 7. CURRENT CONSERVATION sis is needed. Ideally, regional maps of species STATUS AND LONG-TERM could be used in a conservation phylogenetics OUTLOOKS approach to identify which regions contain the most critical species for conservation. There 7.1. Africa are no current initiatives to assess the conservation status of Malagasy bee species. How- For Africa there is an urgent need to iden- ever, because of the extent to which ecosystem tify areas with a high bee diversity and en- degradation has occurred, all areas containing demicity, and conservation strategies for these remnant biota require protection – such areas areas need to be specifically designed and pri- are now very limited and rapidly disappearing. oritised. Nothing like this exists in Africa and Our field work (Schwarz, unpubl. data) found it will not be easy to develop. With the re- that even small pockets of remnant vegetation cent awareness of the importance of pollina- can still support high bee diversity, but many tion services, Africa has come up with a plan of these are still being destroyed by slash-burn Bee conservation in Africa and Madagascar 363 incursions. Expanding such remnants and link- Much of what is now required involves building them via habitat corridors is sorely needed, ing local knowledge infrastructure, which is but the most immediate challenge is to simply needed to promote and engage communities prevent the existing pockets from disappearing and government in conservation. We suggest altogether. the following as some important initial steps: In summary we do not have enough data to thoroughly document bee diversity in Africa • Public awareness needs to be increased. and Madagascar, let alone develop prioritised Campaigns, such as the North American conservation strategies. Furthermore, at the Pollinator Protection Campaign (NAPPC governmental and funding-agency levels, bee – http://www.nappc.org) could be devel- conservation will have to compete with more oped and used to engage farmers, conser- ‘glamorous’ taxa, such as mammals, reptiles vationists, gardeners and other land users. and birds, where conservation measures will A number of these are given in Eardley hold greater cachet with the public. If strate- et al. (2006), but additional case exam- gies for conserving bee diversity in Africa and ples are needed and different media should Madagascar are to be developed and imple- be used for publication. Land users should mented by governments, they will require per- be made aware of comparable situations suasive arguments that bees are important for to theirs where pollinators have been pro- broad ecosystem health and conservation. If tected and encouraged to implement simi- conservation strategies are to be implemented lar measures. at a local level, then those people who decide • Networks, such as API, should be ex- how their land is managed will need to be con- panded to create an active community that vinced that bee-specific measures will be to will enable people interested in bees to their advantage. communicate with one another. These can At a broad level, we require bee diver- also be used to coordinate wider participa- sity surveys across the continent, document- tion in surveys and research projects. ing of food plants and pollinated plants, know- • Much effective research results from the ing the identity and phylogenetic relationships effect of ‘champions’ who promote their between bee taxa, understanding their biol- subjects. But many potential champions do ogy and their interaction with other organ- not develop due to a lack of training and isms. This calls for increased capacity build- mentorship. Support networks are needed ing at all levels. Taxonomic impediments and to develop emerging champions in differ- a dearth of well-curated local collections are ent institutes and countries. major barriers to pollination ecology and pol- • Many Northern Hemisphere-based scien- linator conservation in Africa. In sub-Saharan tists work in Africa and mentor develop- Africa, there are only a few countries, such as ing scientists. Local mentors, however, are South Africa, Kenya and Ghana, that have sub- needed and overseas researchers can help stantial capacity and enthusiasm for pollinator them develop. conservation. There are many more countries • Taxonomists need collections and cata- without a single bee taxonomist or pollina- logued literature. The latter should soon be tion ecologist. In order to overcome these chal- available in Eardley and Urban (in prepara- lenges, there is a need for local bee inventories, tion), but developing collections is a long, regional courses and increased collaboration slow process. The problem could be allevi- between countries. ated through databases of African material in overseas collections and virtual museums (good quality photographs on the In- 8. WHERE TO NOW? ternet) of type material. But the development of local collections is also essential. The challenges that face bee biodiversity • Existing politically stable regions are conservation in Africa are not insurmountable, best placed to spearhead conservation but the pace of development is far too slow. initiatives and these could involve the 364 C.D. Eardley et al.

establishment of research sites and asso- abeille / Apoidea / protection / biodiversité / ciated stations, with local institutes taking Afrique / Madagascar a lead to attract researchers to such sites. Research and conservation agencies could Zusammenfassung – Erhaltung der Bienen in also ‘adopt’ such research centres. Afrika südlich der Sahara und in Madagaskar: • There is a strong need for research into Diversität, Status und Bedrohungen. Das südlich the effect of non-Apis bees on local agri- der Sahara gelegene Afrika und Madagaskar be- culture and how bee diversity can be en- herbergen einen hohen Reichtum an Bienendiver- couraged in areas outside of conservation sität. Diese Diversität ist aus mehreren Gründen unabdingbar. Bienen sind wichtige Bestäuber sowohl parks. Adoption of bee-friendly practices für landwirtschaftliche Nutzpflanzen als auch in na- will be highest when communities and türlichen Ökosystemen, und sie ermöglichen Ein- governments are aware of the benefits to sichten in verschiedenste Aspekte der evolutionären them. Studies such as those of Kremen Biologie. Neuere Untersuchungen legen nahe, dass et al. (2002, 2004) are invaluable in un- Bienen zunächst in der afrikanischen Region ent- standen sind, hierdurch ist ihre Erhaltung für zu- derstanding the principles that need to be künftige Studien über die frühe Evolution der Bie- investigated in Africa and Madagascar. nen unabdingbar. Allerdings stehen das Verständnis dieser Diversität und die Entwicklung von Strategi- en für ihre Erhaltung vor einigen ernsthaften Her- 9. CONCLUSIONS ausforderungen. Die erste Herausforderung ist die vollständige Dokumentation der Bienendiversität in diesen Regionen. Derzeitige systematische Studi- Africa and Madagascar together contain a en belegen eine eher mittelmäßige Artenvielfalt im wealth of biodiversity. These regions are espe- Afrika südlich der Sahara. Dies könnte aber, insbe- cially important for bees, given a likely origin sondere in den ärmeren Gegenden, auf unzureichen- de taxonomische Untersuchungen zurückzuführen of bees in Africa (Danforth et al., 2006). The sein. Obwohl eine neuere Übersichtsstudie über die loss of bee species in these regions will per- gesamte Fauna von Madagaskar durchgeführt wur- manently erode this window into a key part of de, gibt es Gründe zu der Annahme, dass noch viele evolutionary history. Yet the centres of grav- weitere Arten beschrieben werden müssen. ity for bee research in these regions have long Es ist generell anerkannt, dass Bienen für die Be- been in North America and Europe. Continu- stäubung von herausragender Bedeutung sind. Dies ff erstreckt sich aber im wesentlichen auf die Ho- ation of these research e orts is needed, but nigbienen und die Stachellosen Bienen, während effective conservation of bees in Africa and die anderen Bienengruppen kaum beachtet werden. Madagascar will only happen with local en- Wenn gegenüber Regierungen und lokalen Kom- gagement, and this requires building local ex- munen ein überzeugendes Argument für ihre Er- haltung vorgebracht werden soll, ist es notwendig pertise and local initiatives. The first steps die von den nicht-Apis Bienen erbrachten ökologi- have been taken with initiatives like the API, schen Leistungen zu dokumentieren. Die Diversi- but these must be strengthened and extended. tät der Bienen ist einer Anzahl ernster Bedrohun- This will require an international effort, but it gen ausgesetzt, insbesondere der Verschlechterung can not be at arms length. The solutions must oder Zerstörung ihrer Habitate. Während es in einigen Ländern ausgedehnte Schutzgebiete wie zum ultimately come ‘Out of Africa’ (and Mada- Beispiel Nationalparks gibt und diese auch extrem gascar). wichtig sind, ist es für große Teile von Afrika oder Madagaskar unangemessen sich auf Erhaltungsge- biete zu stützen, in denen die Landwirtschaft aus- ACKNOWLEDGEMENTS geschlossen wird. Stattdessen sind für die Erhaltung der Bienendiversität in Gegenden außerhalb von Er- haltungsgebieten die öffentliche Erziehung und die We thank Ms T. Perregil of the Transvaal Mu- Entwicklung landwirtschaftlicher Arbeitsverfahren seum for assistance in finding references, and sowie Unternehmungen wie Ökotourismus zu för- Trevor Lehmeyer for producing the figures. dern. Die Dokumentation der Bienendiversität in Afrika und Madagaskar wird durch den Mangel an Bienen- Conservation des abeilles en Afrique sub- biologen in diesen Gegenden sowie durch schlechte saharienne et à Madagascar : diversité, statut et Infrastruktur und begrenzte Zugangsmöglichkeiten menaces. zu wissenschaftlichen Arbeitsmitteln wir etwa Bee conservation in Africa and Madagascar 365

Büchereien erschwert. Trotz der Einschränkungen Gauld I.D. (Eds.), Hymenoptera and Biodiversity, ist die Entwicklung lokaler Bienensachkenntnis und CAB International, Wallingford, p. 348. die Ausbildung lokaler Forschungs- ‘Champions’, Gess F.W., Gess S.K. (2003) A catalogue of flower vis- die jüngere Biologen führen und anleiten können, iting records for aculeate wasps and bees in the sehr wichtig. Hierbei können die entwickelten Län- semi-arid to arid areas of southern Africa, Albany der helfen. Obwohl Finanzierungsagenturen und Museum, Grahamstown, p. 529. Wissenschaftler aus entwickelten Ländern generell bei der Erhaltung der Bienen helfen können, ist Gikungu M.W. (2006) Bee Diversity and Some es letztendlich am wichtigsteng, lokale Möglichkei- Aspects of Their Ecological Interactions with ten wie Museumssammlungen, Zugang zu Informa- Plants in a Successional Tropical Community, tionsquellen und eine Grundlage an fähigen For- PhD thesis, University of Bonn, Germany. schern aufzubauen. Gikungu M.W., Njoroge G. (2007) Species Diversity, Plant Resources and Conservation Status of Bienen / Erhaltung / Biodiversität / Afrika / Ma- Stingless Bees in Mwingi District, Eastern Kenya: dagaskar / Apoidea Discovery and Innovation (RPSUD Special Edition), Vol. 19, No. 3. Gikungu M.W., Winfree R., Gemmill-Herren B. (2008) Agricultural Ecosystems May Support High Levels of Pollinator Diversity and REFERENCES Abundance (Poster Paper), 13th SBSTTA Meeting of the Convention on Biological Diversity, 18-22 Byrne A., Fitzpatrick Ú. (2009) Bee conservation pol- February 2008, Rome, Italy, CBD Technical icy at the global, regional and national levels, Series No. 34, pp. 39–41. Apidologie 40, 194–210. Hepburn H.R., Guye S.G. (1993) An Annotated bib- Chenoweth L.B., Schwarz M.P. (2007) social biology liography of the Cape honeybee, Apis mellif- of two Malagasy Halterapis: evidence that euso- era capensis Eschscholtz (Hymenoptera: Apidae), ciality is plesiomorphic for an ancestral allodapine Afr. Entomol. 1, 235–252. lineage, Ann. Entomol. Soc. Am. 100, 311–319. Johannsmeier M.F. (2001) Beekeeping in South Chenoweth L.B., Schwarz M.P., Fuller S.F., Tierney Africa, Plant Protection Research Institute, S.M., Park Y.C. (2008) Hasinamelissa:anew Handbook No. 14, Plant Protection Research genus of allodapine bees from Madagascar re- Institute, Agricultural Research Council, Pretoria. vealed by DNA sequence data and larval morphology, Syst. Entomol. 33, 700–710. Koch C. (1958) Preliminary notes on the coleoptero- Danforth B.N., Sipes S., Fang J., Brady S.G. (2006) logical aspect of the arrow poison of the bushmen, The history of early bee diversification based on Pamphlet S. Afr. Biol. Soc. 20, 2–41. five genes plus morphology, Proc. Natl. Acad. Sci. Kremen C., Williams N.M., Thorp R.W. (2002) Crop USA 103, 15118–15123. pollination from native bees at risk from agricul- Danforth B.N., Eardley C., Packer L., Walker K., Pauly tural intensification, Proc. Natl. Acad. Sci. USA A., Randrianambinintsoa F.J. (2008) Phylogeny of 99, 16812–16816. Halictidae with an emphasis on endemic African Kremen C., Williams N.M., Bugg R.L., Fay J.P., Thorp Halictinae, Apidologie 39, 86–101. R.W. (2004) The area requirements of an ecosys- Eardley C.D. (2003) Revision of the Afrotropical tem service: crop pollination by native bee com- Ctenoplectrini (Hymenoptera: Apidae), African munities in California, Ecol. Lett. 7, 1109–1119. Plant Protection 9, 5–18. Kuhlmann M. (2005) Diversity, distribution pat- Eardley C.D., Roth J., Clarke S., Buchmann S., terns and endemism of southern African bees Gemmill B. (2006) Pollinators and Pollination: (Hymenoptera: Apoidea), in: Huber B.A. (Ed.), A resources book for policy and practice, API, African Biodiversity, Springer, Netherlands, Pretoria, 77 p. pp. 167–172. Engel M.S. (2001) A monograph of the Baltic amber McCall R.A. (1997) Implications of recent geological bees and evolution of the Apoidea, Bull. Am. Mus. investigations of the Mozambique Channel for the Nat. Hist. 259, 1–192. mammalian colonization of Madagascar, Proc. R. Fuller S., Tierney S.M., Schwarz M.P. (2005) Soc. Lond. B 264, 663–665. Phylogenetics of the allodapine bee genus Martins D., Gemmill B., Eardley C.D., Kinuthia Braunsapis: implications for across sea dispersal W., Kwapong P., Gordon I. (2003) Plan of of bees, J. Biogeogr. 32, 2135–2144. Action of the African Pollinator Initiative, Gess F.W., Gess S.K. (1993) Effects of Increasing African Pollinator Initiative, Nairobi, republ. Land Utilization on Species Representation and in 2007 [online] ftp://ftp.fao.org/docrep/fao/010/ Diversity of Aculeate Wasps and Bees in the Semi- a1490e/a1490e00.pdf (accessed on 5 February Arid Areas of Southern Africa, in: LaSalle J., 2009). 366 C.D. Eardley et al.

Michener C.D. (1975) A taxonomic study of African Madagascar: Implications for ancient and modern allodapine bees, Bull. Am. Mus. Nat. Hist. 155, faunas, J. Geol. 108, 275–301. 67–240. Schwarz M.P., Fuller S., Tierney S.M., Cooper S.J.B. Michener C.D. (1981) Classiﬁcation of the Bee (2006) Molecular phylogenetics of the exoneurine Family Melittidae with a Review of Species of allodapine bees reveal an ancient and puzzling dis- Meganomiinae, Contrib. Am. Entomol. Inst. 18, persal from Africa to Australia, Syst. Biol. 55, 31– 1–135. 45. Michener C.D. (2007) The Bees of the World, The Shaw E.M., Wooley P.L., Rae F.A. (1963) Bushman John Hopkins University Press, Baltimore and arrow poisons, Cimbebasia 7, 2–41. London, 913 p. Smith J.A., Tierney S.M., Park Y.C., Fuller S., Schwarz Myers N., Mittermeier R.A., Mittermeier C.G., da M.P. (2007) Origins of social parasitism: The im- Fonseca G.A.B., Kent J. (2000) Biodiversity portance of divergence ages in phylogenetic stud- hotspots for conservation priorities, Nature 403, ies, Mol. Phylogenet. Evol. 43, 1131–1137. 853–858. Stringer C., Andrews P. (2005) The Complete World of Pasteels J.J. (1965) Révision des Megachilinae Human Evolution, Thames & Hudson, London. de l’Afrique Noire: les genres Creightoniella, Chalicodoma et Megachile (s.str.),Ann.Mus.R. Whitehead V.B., Steiner K.E. (2001) Oil-collecting Afr. Centr. 137, 1–563. bees of the winter rainfall area of South Africa Pauly A., Brooks R.W., Nilsson L.A., Persenko Y.A., (Melittidae, Rediviva), Ann. S. Afr. Mus. 108, Eardley C.D., Terzo M., Griswold T., Schwarz 143–277. M., Patiny S., Munzinger J., Barbier Y. (2001) Whitehead V.B., Schelpe E.A.C.L.E., Anthony N.C. Hymenoptera Apoidea de Madagascar, Ann. Zool. (1984) The bee, Rediviva longimanus Michener Wet. 128, 1–390. (Apoidea: Melittidae), collecting pollen and Rasmussen C., Cameron S.A. (2007) A molecular phy- oil from Diascia longicornis (Thumb.) Druce logeny of the Old World stingless bees and the (Scrophulariaceae), S. Afr. J. Sci. 80, 286. non-monophyly of the large genus Trigona, Syst. Whitehead V.B., Steiner K.E., Eardley C.D. (2008) Oil Entomol. 32, 26–39. collecting bees mostly of the summer rainfall area Reyes S.G. (1991) Revision of the bee genus of southern Africa (Melittidae, Rediviva Friese), J. Braunsapis in the Oriental region, Univ. Kans. Sci. Kans. Entomol. Soc. 81, 122–141. Bull. 54, 179–207. Yoder A.D., Nowak M.D. (2006) Has vicariance or dis- Rogers R.R., Hartman J.H., Krause D.W. (2003) persal been the predominant biogeographic force Stratigraphic analysis of Upper Cretaceous in Madagascar? Only time will tell, Annu. Rev. rocks in the Mahajanga Basin, northwestern Ecol. Evol. Syst. 37, 405–431.