Few regional or continent-wide assessments of bird use for traditional medicine have been attempted anywhere in the world. Africa has the highest known diversity of bird species used for this purpose. This study assesses the vulnerability of 354 bird species used for traditional medicine in 25 African countries, from 205 genera, 70 families, and 25 orders. The orders most represented were Passeriformes (107 species), Falconiformes (45 species), and Coraciiformes (24 species), and the families Accipitridae (37 species), Ardeidae (15 species), and Bucerotidae (12 species). The Barn owl (Tyto alba) was the most widely sold species (seven countries). The similarity of avifaunal orders traded is high (analogous to “morphospecies”, and using Sørensen's index), which suggests opportunities for a common understanding of cultural factors driving demand. The highest similarity was between bird orders sold in markets of Benin vs. Burkina Faso (90%), but even bird orders sold in two geographically separated countries (Benin vs. South Africa and Nigeria vs. South Africa) were 87% and 81% similar, respectively. Rabinowitz's “7 forms of rarity” model, used to group species according to commonness or rarity, indicated that 24% of traded bird species are very common, locally abundant in several habitats, and occur over a large geographical area, but 10% are rare, occur in low numbers in specific habitats, and over a small geographical area. The order with the highest proportion of rare species was the Musophagiformes. An analysis of species mass (as a proxy for size) indicated that large and/or conspicuous species tend to be targeted by harvesters for the traditional medicine trade. Furthermore, based on cluster analyses for species groups of similar risk, vultures, hornbills, and other large avifauna, such as bustards, are most threatened by selective harvesting and should be prioritised for conservation action.
Citation:Williams VL, Cunningham AB, Kemp AC, Bruyns RK (2014) Risks to Birds Traded for African Traditional Medicine: A Quantitative Assessment. PLoS ONE 9(8): e105397. doi:10.1371/journal.pone.0105397
Editor: David L. Roberts, University of Kent, United Kingdom
Received: March 16, 2014; Accepted: July 23, 2014; Published: August 27, 2014
Copyright: © 2014 Williams et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability:The authors confirm that all data underlying the findings are fully available without restriction. Most relevant data within the paper are within the paper and its Supporting Information files. Further data on population size, extent of occurrence (EOO), number of habitats, and average body mass per species are available from the BirdLife website www.birdlife.org/datazone/species/search
Funding:VLW was funded by the University of the Witwatersrand SPARC Prestigious and URC Postdoctoral Fellowships. RKB acknowledges funding from the National Research Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
The balance between culture, ritual, commerce and conservation is an emotive issue, particularly where charismatic animal species are used for traditional medicine (TM). For example, the expanding trade in animal derivatives, such as tiger bones, bear gallbladders and rhino horns in the Far East for Traditional Asian Medicine, is especially controversial and of international concern, since the species concerned are endangered yet in high demand . The practice of healing using animals (‘zootherapy’) –, however, has deep historical origins. Civilizations in Ancient Egypt, Mesopotamia and China have written records of therapies that require bat limbs, mongoose blood or glands from musk deer, respectively, and included in these archives from a bygone age are medicinal remedies using swallow's liver, bird excrement and chicken eggs .
Indigenous knowledge (IK) related to the consumptive utilisation of avifauna span continents and cultures. At a global scale, at least 4,173 bird species (42% of 9,856 extant avian species) are used by people, mainly as pets (37%) or food (14%), with far fewer used in sport hunting, ornamentation or TM (<1%, see  Figure 1). In the early 1900s, various authors (e.g. –) documented customary uses and reverence for birds within indigenous cultures (e.g. “Woe betide the native who…kills one of these birds; he will be struck down by…illness, which…will terminate in his death” ), which stemmed from curiosity about ethno-ornithology and the indigenous use of avifauna. In more recent years, concerns have emerged about both the sustainability of utilisation and the need to document IK in the face of its apparent erosion due to modernization and adoption of western medicinal and cultural practices. Accordingly, researchers have turned their attention to inventorying species and documenting zootherapeutic practices , –.
number of species across 25 African countries.
Growth in the number and size of markets for TM has been correlated with increased numbers and levels of species harvested . As TM markets have grown, so has commercial availability of targeted taxa and concerns that preferred species are being acquired in an unsustainable manner. However, few studies have quantified the trade or evaluated the impact that unregulated commercial collection may be having on populations of threatened avian species. In Africa, most studies conducted to date are country-specific surveys of TM or fetish markets, or taxon-specific studies focussed on such birds as ground-hornbills, owls or vultures. Cocker & Mikkola , for example, suggested that harvesting may involve “thousands, possibly tens of thousands, of owls annually”. Bruyns et al.  reported on the sale of Southern Ground-hornbill (Bucorvus leadbeateri) at a market in Zimbabwe and concluded that since hunting for this species was mainly opportunistic, the medicinal trade was not likely to be a significant threat to its national population. National surveys have been carried out in Benin , Nigeria , , and South Africa –. One of the most comprehensive country surveys is of the quantities of bird species observed in each of 24 Nigerian markets, together with insights into how species are collected, priced and used, and anthropogenic factors that directly or indirectly influence their trade .
In response to the growing TM trade as a potential threat to wildlife in South Africa, Cunningham & Zondi  conducted one of the first ethno-zoological trade assessments in 1991. Vultures, Bateleur Eagles (Terathopius ecaudatus), and Southern Ground-hornbills were identified as conservation priorities. Concerns regarding the hunting of vultures for TM in South Africa, especially the near-endemic and threatened Cape Vulture (Gyps coprotheres), led Mander et al.  to quantitatively assess the vulture trade in South Africa, consulting traditional healers and making recommendations for conservation action. This, and Cocker & Mikkola's work on owls, identified priority conservation taxa in the TM trade , .
In contrast to these national studies, even fewer regional or continent-wide assessments have been conducted in Africa. Only Marshall's  report on medicinal wildlife resources in East and southern Africa incorporated a wide-ranging account of birds in TM and their associated use, as well as their trade and conservation statuses within the source countries. More recently, Williams et al.  (the authors of this paper) quantitatively assessed the richness and rarity levels of avifauna used and sold for TM within 25 African countries. In assessing rarity, the study was partly reliant on species-specific data published on the BirdLife International website [www.birdlife.org]. However, the 2012 update of the IUCN Red List for birds resulted in new and revised information being available for a large proportion of the species investigated, thereby rendering the book chapter out-dated by the time it was published. The current paper re-assesses the vulnerability of these avian taxa to harvesting on the basis of these more recent data, but also includes new analyses to account more comprehensively for the richness and prevalence of avian taxa sold in African TM markets, substantially updating Williams et al. . Our objectives were to: i) update the list of avian species recorded in TM markets using published accounts and personal observations; ii) re-examine patterns of rarity and commonness among avifauna used; iii) relate mean body size (mass) to inherent rarity and the prevalence of species in the markets, and iv) detect taxa that may be vulnerable to selective harvesting, in conjunction with current and potential threats to their existence, and which may arise from any escalation in their use and commercial harvesting.
The inventory of birds recorded in TM markets and shops in Africa (Figure 1) was compiled from published accounts , –, –, and supplemented by our own personal research, observations and photographs taken in various TM markets across Africa over 23 years (1989–2012). The most comprehensive published information available was for birds sold in TM markets in Benin (BJ) , Nigeria (NG) , and South Africa (ZA) . In addition, market data for BJ (Dantokpa market), Burkina Faso (BF; Ouagadougou), Côte D'Ivoire (CI; Abidjan and Bouake), Togo (TG; Lome market), ZA (Johannesburg) and Zimbabwe (ZW; Bulawayo) were supplemented with identifications made from photographs taken during fieldwork (Figure 2). However, we do not consider our inventory to be complete for Africa since the information was patchy and there was a paucity and/or absence of information for certain regions, especially East Africa (Figure 1). Hence, some of the findings must be viewed in light of these information gaps. The complete inventory for 399 taxa is published in Table S1.
A. Heads of a variety of species, including a Western Grey Plantain-eater (Crinifer piscator), Double-toothed Barbet (Lybius bidentatus), Red-billed Hornbill (Tockus erythrorhynchus) and Double-spurred Francolin (Francolinus albogularis) (Ouagadougou market, Burkina Faso). B. Vultures and raptors, a high conservation priority group, sold here in Xipamanine market, Maputo, Mozambique. C. A basket of more than 15 species, including Broad-billed Roller (Eurystomis glaucurus), Fine-spotted Woodpecker (Campethera punctuligera), African Wood-Owl (Strix woodfordii), Rose-ringed Parakeet (Psittacula krameri), White Helmet-Shrike (Prionops plumatus), and Standard-winged Nightjar (Macrodipteryx longipennis) (Dantokpa market, Benin). D. Senegal Parrot (Poicephalus senegalus) (Ouagadougou market, Burkina Faso). [Photos: A.B. Cunningham]
Ornithological Classification and Enumeration
The classification and nomenclature of BirdLife International was followed because their taxonomic list is kept current and forms the basis for the IUCN Red List assessments (along with Tobias et al.  for the forthcoming checklist of the birds of the world ). Furthermore, it was important that the quality of quantitative information obtained for the majority of species was consistent and derived primarily from the same source. While the BirdLife ornithological classification differs slightly from other taxonomic lists available online, it was necessary to adhere to one system. This meant, however, that five taxa considered separate species in some ethno-avian literature were ‘lumped’ with other species for this paper, namely: (i) Burchell's Coucal (Centropus burchellii) with White-browed Coucal (Centropus superciliosus) (ZA, BJ); (ii) Dark-capped Bulbul (Pycnonotus tricolor) with Common Bulbul (Pycnonotus barbatus) (BJ, NG, ZA); (iii) Sahel Paradise-Whydah (Vidua orientalis) with Eastern Paradise-Whydah (Vidua paradisaea) (NG); (iv) Cape White-eye (Zosterops virens) with Pale White-eye (Zosterops pallidus) (ZA); and (v) African Hoopoe (Upupa africana) with Eurasian Hoopoe (Upupa epops) (BF, NG, SD, ZA and Morocco, MA), although the BirdLife Taxonomic Working Group is reviewing the latter treatment. Taxonomic data and IUCN Red-Listing are correct to June 2014.
We were conservative in our enumeration of the total number of avian taxa. If a bird could not be recognized beyond genus, it was not included in any further analyses. Also excluded, except where specified, were:
- Two exotic species recorded in the markets, namely Indian Peafowl (Pavo cristatus) and Common Myna (Acridotheres tristis) since, while these species might be used now, they would not have formed part of traditional inventories;
- 15 species of Palearctic (PAL) non-breeding migrants to Africa (Common Cuckoo, Spotted Flycatcher, Montagu's Harrier, Pallid Harrier, Northern House-martin, Red-necked nightjar, Northern Pintail, Kentish Plover, European Roller, Great Snipe, Jack Snipe, White Stork, Barn Swallow, Yellow Wagtail, Eurasian Wryneck) (respectively, Cuculus canorus, Muscicapa striata, Circus pygargus, Circus macrourus, Delichon urbicum, Caprimulgus ruficollis, Anas acuta, Charadrius alexandrinus, Coracias garrulus, Lymnocryptes minimus, Ciconia ciconia, Hirundo rustica, Motacilla flava, Jynx torquilla). The PALs were omitted because much of the data used to assess threats to African species would not apply (even for the few species with a small breeding range extending into the Palearctic extremities of North Africa or into southern Africa), complicated further by estimating the extent of their actual wintering range in Africa versus areas visited in transit to and from the Palearctic summer breeding range. Note that there are other intra-African migrant species with separate breeding and non-breeding ranges but, since both ranges fall within the Afrotropics, they are not considered separately from other Afrotropical species. However, given the substantial declines that PALs have undergone in the last 30 years , , we have discussed the findings for PALs separately;
- 49 species recorded as being used for TM but not recorded in the TM markets, since our analyses are based on species selectively harvested or acquired for the commercial trade.
The richness and percentage-similarity of species and orders sold in markets in various African countries were compared using the Sørenson Index (for incidence-based data), calculated using EstimateS (version 7.5.1) .
Patterns of rarity and commonness
One way of examining avian vulnerability to consumptive use is to classify the species based on the probability of them becoming rare if exploitation and persistent, selective, commercial hunting become deterministic factors that threaten population dynamics. Rabinowitz developed the ‘seven forms of rarity’ model that was originally applied to assess the vulnerability of plants , . The model was based on three variables that indicated the level of rarity, namely: range size, habitat specificity and local abundance (Table 1). When species are dichotomized for each of the variables, the result is an eight-cell model (Table 1) that Yu & Dobson  adapted and used to create four ranks of rarity for assessing the rarity and commonness of mammals. For example, Category H species are rare in all three factors and assigned a rank of 1, whereas Category A species are common and assigned a rank of 4.
Yu & Dobson's  classes of rarity were applied to the data by placing each species in a category ranging from A to H. Thereafter, the categories were assigned ranks from 1 to 4 (most to least rare respectively; Table 1). The purpose was to examine patterns of rarity and commonness, comparing these classifications quantitatively with other variables to detect taxa that may be vulnerable to the TM trade. The data used to do the rarity assessments were mostly obtained from the BirdLife International website (http://www.birdlife.org/datazone/species), including: (i) estimates of population size, (ii) population trends (i.e. increasing, decreasing, stable), (iii) Extent of Occurrence (EOO, km2), (iv) number of Level 1 habitats, and (v) number of African countries in which the species occur. Some habitat and population abundance (‘status’) assessments were validated against information obtained from Sinclair & Ryan . All these data were correct as of December 2012.
Assigning species to the rarity categories requires dichotomizing the distribution (large or small), habitat (broad or narrow) and population abundance (large/high/dominant or small/rare/non-dominant). For the distribution range, the median EOO for all the inventoried species was determined to be 6,790,000 km2 (range: 59,500 km2 to 63,300,000 km2; n = 346) and therefore EOOs greater and smaller than the median were considered to be large and small ranges respectively. Dichotomizing habitat and abundance were more subjective, since habitat types are essentially various graded combinations of geological, topographical, climatic, aquatic and vegetation features, ranging from desert to rainforest, and abundance ranges from rare to abundant. Sinclair & Ryan  were used to assist in borderline judgments.
One-way ANOVAs were computed to test the differences in the mean EOOs for species sold in markets that were assigned to the rarity groups ranked from 1 (most rare) to 4 (most common). The significance between group means was determined using the post-hoc Tukey HSD multiple comparison test for unequal N using Statistica 6 (Statsoft Inc).
IUCN Red List Status
The IUCN status and threats to avian diversity were further examined based on the statuses available on the IUCN Red List of Threatened Species (www.iucnredlist.org, downloaded December 2012 and June 2014) and BirdLife International websites. The Red List status of a species will indicate whether they are likely to be increasingly threatened if continuous high-impact harvesting occurs.
Body Size and Mass
Given evidence that extinction risks to birds incurred through human persecution was correlated with large body size , we calculated and compared the mean mass of traded birds in the different rarity classes and IUCN threat categories, as well as birds specifically sold at markets in ZA, BJ and NG. Mean body mass was obtained for 344 species from the BirdLife International website and for southern Africa from Chittendon & Upfold . Mass was used as a proxy for size because mass is used in scaling analyses that predict that larger birds will occur at lower densities, have lower recruitment rates, live longer and therefore be more vulnerable to harvesting . The average mass of the Ostrich (Struthio camelus) (114 kg) was excluded from all analyses since the species is a statistical outlier by more than three standard deviations when regressed against body length (r = 0.51 for n = 203 including Ostrich; r = 0.84, p<0.001, n = 202 excluding Ostrich).
Cluster Analysis and Conservation Priorities
Cluster analysis is an effective way of identifying groups of species with profiles of similar risk and/or conservation priority in relation to criteria chosen to characterise these risks –. One purpose of a cluster analysis is to partition objects (such as species) into groups suggested by the data rather than defined a priori, so that objects in a given cluster tend to be relatively similar to each other and objects in different clusters are dissimilar . In ethno-ecological studies, determinations of conservation priorities are regularly made using linear numerical rating systems, whereby ranked values are assigned to species for the different variables chosen. This method imposes artificial linearity onto a naturally non-linear system, tends to rank many species together and makes objective separation of species into priority hierarchies difficult . Since the selection of hierarchical boundaries is often subjective or arbitrary, a recommended approach is to generate importance values based on ranks assigned within variables, sorting species in order of the variable values (or, total score) to place similar species nearby in ‘multivariate space’ , and using this order when conducting a cluster analysis .
K-means clustering is a simple non-hierarchical classification method appropriate as a data-reduction technique where there are large numbers of species, no real dependent variable and it is desirable to determine whether groups of similar species exist , . Since this study is based on an inventory that we do not consider to be complete, and it was not possible to collect data for all the potential variables for all species, we selected only three variables for which we could obtain data for as many species as possible (280 of the 306 species recorded in the markets) so as to make a generalised assessment of priority and vulnerability.
The variables selected for the cluster analyses were: i) the four ranks assigned to the Rabinowitz rarity classes (since the system classifies species based on habitat specificity in addition to range and population size) (after Table 1); ii) mean body mass (since size was found to be related to vulnerability); and iii) the number of countries that recorded a species in at least one market (this, however, depended on the number of studies from which we could extract information). Since the variables were not all measured on the same numerical scale (e.g. body mass is continuous, and number of markets is discontinuous), the variable values were standardized and similarly scaled. The data were converted to scores per variable of between 0 (the lowest score and least vulnerable) and 1 (the highest score and most vulnerable). To do this, the values for each species in a corresponding variable for body mass and number of markets was divided by the highest value in that column but, for the rarity rank, the values first had to be reversed before being standardised since the original rank of 4 implied commonness and not rarity. The total scores for the three variables were summed, and the species arranged in descending order from highest to lowest scores (maximum score = three). Statistica 6 was used to perform the K-means cluster analyses. The process was repeated by progressively specifying the formation of two to four clusters and then evaluating the species within each cluster, ending up with only two groups of ‘higher’ (S = 115) and ‘lower’ (S = 165) conservation priorities and vulnerability (Table S2).
We recorded 399 avian taxa as being used and sold for TM in 25 African countries (Tables 2 & S1), of which 354 were identified to species (Table 2). The species were from 207 genera, 70 families and 25 orders (Tables 3). When 49 species not recorded in the markets were excluded from the list, then the total ornithological richness of birds sold in the markets was 306 species from 189 genera, 69 families and 25 orders (including PALS and exotics) from 7 African countries (BF, BJ, CI, NG, TG, ZA and ZW) (Table 3).
Perching songbirds (Passeriformes), which comprise 56% of Africa's bird species (inferred from BirdLife International website, 2013), had the highest number of recorded taxa in use (23 families, 61 genera, 107 species) and in trade (50 genera, 79 species) (Table 3). Within the traded Passeriformes, the Sturnidae (starlings) was the most prevalent family (9 species). Of all the families in TM trade, the Accipitridae (Order Falconiformes) had the most recorded genera (26 genera; 37 species; including kites, hawks, eagles, vultures), followed by the Ardeidae (Order Ciconiiformes) (11 genera; 15 species; including herons and egrets) (Table S1). The Bucerotidae (hornbills), Cuculidae (cuckoos) and Strigidae (owls) were the next most specious families in trade (12, 11 and 11 species per family, respectively).
The highest number of species were recorded in NG markets (200 species, plus one exotic, five PAL and six unidentified species), followed by BJ (134 species, plus nine PAL and nine unidentified), ZA (84 species, plus two exotic and 24 unidentified), BF (29 species, plus one PAL and 11 unidentified), CI (12 species, plus three unidentified), TG (11 species, plus 13 unidentified) and ZW (six species, plus six unidentified) (Figure 1). Fourteen PAL species were only recorded in NG, BJ and BF markets. The Red-necked Nightjar was only recorded as being used (not traded) in MA. Five intra-African migrants were also identified (Wahlberg's Eagle, Bronze-winged Courser, Levaillant's Cuckoo, Violet-backed Starling, African Golden Oriole; respectively Aquila wahlbergi, Rhinoptilus chalcopterus, Clamator levaillantii, Cinnyricinclus leucogaster, Oriolus auratus), but only in NG and BJ markets, despite their total range in Africa extending to southern Africa.
Frequency and Similarity of Species Traded in Markets
Ostriches were the most commonly used of all the birds (12 countries), but the species has only been positively identified in the markets of four countries to date (MA, NG, ZA, ZW) (Table 4). The most frequently recorded species in markets were the Barn Owl (seven countries), Pied Crow (Corvus albus; six countries), and Hooded Vulture, Helmeted Guinea Fowl and African Pied Hornbill (Necrosyrtes monachus, Numida meleagris, Tockus fasciatus, respectively; five countries each) (Table 4). Of the top 19 species, only three are currently threatened and the populations of six species are experiencing declines in numbers (Table 4). Fifteen of the top 19 species are, however, widespread with EOOs >7 million km2. The EOO of the Barn Owl is estimated to be 63 million km2 , making it one of the species most likely to be selected for TM use given its cultural significance and spatial commonness. The African Pied Hornbill, Western Grey Plantain-eater (Crinifer piscator) and Grey Parrot (Psittacus erithacus), despite having ranges restricted to <4.5 million km2 and occurring in specific habitats, are locally abundant species frequently available in West African markets.
The countries most similar in terms of the species sold in the markets are BJ vs. NG (53% similar in terms of Sørenson's index; 93 species in common), followed by NG vs. ZA (27% similar, 39 species in common) and BJ vs. ZA (24% similar, 28 species in common) (Table 5). BJ, NG and ZA also have 23 species in common between them. Despite the limited survey work conducted in BF, the species sold in the markets there are 24% similar to those sold in BJ.
While there are geographical differences in the occurrence of species, and hence their occurrence in markets, the selection of birds for medicine is often at a less specific ‘morphospecies’ level (i.e. a typological 'species' that can only be identified as owl, vulture or kingfisher). In South Africa, for example, all vultures, regardless of species, are generically referred to in isiZulu as ‘iNqe’. When re-assessing the frequencies of birds in the markets of seven countries at the level of order (analogous to morphospecies), the most prevalent morphospecies identified were: 1) owls (Strigiformes), hornbills (Bucerotiformes), diurnal birds of prey (Falconiformes), perching birds (Passeriformes) and gamebirds (Galliformes) (seven countries each); 2) storks, herons and egrets (Ciconiiformes) and kingfishers (Coraciiformes) (six countries each); and 3) turacos (Musophagiformes), woodpeckers and relatives (Piciformes), doves and pigeons (Columbiformes), pelicans and relatives (Pelicaniformes) cranes and bustards (Gruiformes) (five countries each). Using Sørenson's similarity index, we also assessed the similarity of avifaunal orders of birds sold in the markets (Table 5). The highest similarity was between bird orders sold in markets of BJ vs. BF (90%), followed by NG vs. BJ (87%). Even bird orders traded in two geographically separated countries (BJ vs. ZA, and NG vs. ZA respectively) were 87% and 81% similar.
Rabinowitz Classification and Patterns of Rarity and Commonness
The number and proportion of species varied within the categories of rarity (Table 6) and within the rarity categories per order (Figure 3). About 23% (69 species) of birds sold in the markets are very common and locally abundant in several habitats over a large geographic area (Category A), whereas 10% (30 species) were in the most rare category and are considered to be consistently sparse in a specific habitat over a small geographic area (Category H) (Table 6).
The number of species per order and the mean rarity rank (derived from rank scores in Table 1) are given in parentheses. The orders are listed from most common to most rare (left to right respectively).
Species within orders were not homogeneously distributed among the categories of commonness and rarity (Figure 3). Compared to other orders, Strigiformes (owls) had the highest proportion of rare species in Category H (23%), followed by Bucerotiformes (hornbills; 21%), and Gruiiformes (cranes; 19%) (Figure 3). Conversely, Columbiformes (doves and pigeons) had the highest proportion of common species (Category A; 78%). While 23% of the traded owls tend to be constantly sparse in specific habitats over a small geographic area (Category H), an equal proportion is locally abundant in specific habitats over a large geographic area (Category B; Figure 3). Hence, when the mean rarity rank per order was calculated (based on Table 1, with ranks closer to 1 indicating greater relative rarity), the orders that were relatively rarer and had a higher proportion of rare species (excluding those with less than five species), were Musophagiformes (mean rank = 1.86; S = 7 species), Bucerotiformes (mean rank = 2.00; S = 14 species), Gruiformes (mean rank = 2.13; S = 16), and Galliformes (gamebirds; mean rank = 2.33; S = 9) (Figure 3). The orders with the highest proportion of common species (i.e. category A) are Columbiformes (mean = 3.67; S = 9), Cuculiformes (cuckoos and relatives; mean = 3.09; S = 11), Coraciiformes (kingfishers and relatives; mean = 3.05; S = 21) and Ciconiiformes (storks; mean = 3.05; S = 19) (Figure 3). The overall mean rank for all traded species was 2.71 (S = 288 species), indicating that there is an intermediate degree of inherent commonness among the species sold for TM across the African continent.
When comparing the observed number of traded species in the eight rarity categories for ZA, NG and BJ (Figure 4), broadly similar proportions of species were allocated to each rarity category. The major differences were, however, that ZA has a smaller proportion of species that are constantly sparse in several habitats over a large geographic area (category C) and a smaller proportion of species that are locally abundant in a specific habitat over a small geographic area (category F) – this is partly because its southern position supports fewer ‘typical’ African habitats, and its range-restricted species are of less favoured types. NG and BJ, however, utilise a larger proportion of species with narrower distribution ranges (categories E to F) – which is partly the result of harvesting species that inhabit the more range-restricted forest and woodland habitats that occur within the harvesting catchment available to the hunters and consumers in those countries.
When examining the individual factors that contribute towards the rarity of avian species sold for TM (Table 7), we found that 45% of species have ‘small’ geographic ranges (i.e. <6,790,000 km2), 24% are non-dominant or constantly sparse within their range, and 60% are quite habitat specific. Chi-squared pair-wise comparisons of the characteristics of rarity indicated range and habitat specificity to be significant factors in the allocation of species to categories (χ2 = 19.7, d.f. = 3, P<0.01). However, habitat specificity in combination with population size were nearly significant factors (χ2 = 7.6, d.f. = 3, P = 0.055), but range and population size combined were not significantly related to categorizations of rarity (χ2 = 1.1, d.f. = 3, P = 0.77).
An EOO of ≤20,000 km2 is the quantitative threshold for classifying a species as threatened, specifically Vulnerable, according to the B1 Red List criterion of the IUCN. However, since birds are rarely range-restricted in the same sense and to the same extent that similarly threatened terrestrial animals and sessile plants are, the smallest range for a traded species in this study was five times larger than the B1 threshold (Knysna Turaco – Tauraco corythaix; 125,000 km2). Hence, the Rabinowitz classification offers an alternative method for evaluating vulnerability where the B1 criterion cannot be applied. The mean EOO of species within each rarity class decreased with increasing species rarity (Table S3). The mean EOO for all traded species was 9,897,230 km2 (S = 283), well above the 6,79 million km2 median for all species considered in this paper (Table S3).
IUCN Red List Status and Population Trends
The IUCN Red List statuses of 17 traded species are threatened (five Endangered; 12 Vulnerable), five are Near Threatened (excluding PALs), and the majority of species (92%) are categorized as Least Concern (Table 6). For the most part, the populations of traded species are stable (51%) and/or increasing (10%; Table 6). However, populations are declining for 30% of species. Major threats to birds vary with species ,  so, in devising conservation strategies, it is important not to focus unnecessarily on the TM trade.
Body Mass and Rarity
From published accounts of birds sold in TM markets throughout Africa, there is evidence to suggest that larger birds have a greater tendency to be selectively harvested and are accordingly more prevalent in the markets , , . Consequently, larger birds are at greater risk of population decline and localised extirpations. When analysing the mean mass of species in the eight rarity classes (excluding the Ostrich in class A), we found that larger birds tended to be non-dominant species (classes C,D,G,H; Figure 5), especially those occurring in specific habitats (classes D and H), whereas smaller birds tended to be locally abundant (classes A,B,E,F; Figure 5). When the factors are dichotomised (Table 8), the largest birds are usually those that occur over a large range (mean mass = 773 g), but have smaller population sizes and/or lower densities (mean mass = 1584 g) and occur in specific habitats (mean mass = 702 g) – the latter two factors increasing the risks of population declines with increased and selective utilisation of larger birds. When comparing the mean body mass of taxa in the adapted four rarity ranks, there is an increase in mean mass with increased rarity: rank 4 (most common) = 520 g±1,078 g (± S.D.); rank 3 = 684 g±1,510 g; rank 2 = 868 g±1,656 g; and rank 1 (most rare) = 1,267 g±2,258 g. Further evidence for larger birds being more vulnerable than smaller birds was derived from the mean mass calculated according to the IUCN Red List categories. Whereas traded species classified as Least Concern weighed on average 464 g±1,035 g (n = 259), threatened (Endangered and Vulnerable) species sold in the markets were 8.3 times heavier (3,832 g±2,598 g, n = 17), and Near Threatened species were 8.8 times heavier (4,083 g±1,592 g, n = 5).
When analysing the mean mass of species sold in ZA, NG and BJ markets, a similar pattern to that in Figure 5 emerged for each country (Figure S1). Furthermore, birds sold in markets in ZA were, on average, larger than those sold in markets in BJ and NG (1,197 g, 657 g, and 618 g respectively). Reasons for the average size of birds being heavier/larger in South Africa compared to West Africa are likely to be related to the trend in the sizes of the species that occupy the forest and woodland habitats within the latter's harvesting and hunting catchment.
Vulnerability to Selective Harvesting
Through cluster analysis, avian species of varying degrees of rarity and commonness were assigned to internally homogenous clusters (groups 1 and 2) that were indicative of their relative conservation priorities (Table S2). These groupings indicated vulnerability to selective harvesting by integrating the biological and ecological traits of species with their frequency in TM markets. Species in Group 1 are generally more vulnerable to selective harvesting since they are inherently larger, rarer birds with populations that tend to be decreasing (Table S2). These species are also less widespread (mean EOO = 4,406,281 km2±3,822,394 km2) and sold in fewer markets than those in Group 2. However, their rarity, rather than consumer preference, is also a factor in them being recorded in fewer markets. This group contained all the Musophagiformes (turacos), many Bucerotiformes (hornbills; 86%), and most of the Galliformes (gamebirds; 67%), Gruiformes (cranes and relatives; 67%) and Psittaciformes (parrots; 60%) (Table S2).
Species in Group 2 are generally less vulnerable to selective harvesting since they tend to be smaller and more common, with populations that are stable or increasing (Table S2). These species have also been recorded in relatively more markets, and have larger ranges (14,037,660 km2±10,816,510 km2). Group 2 contains most of the Columbiformes (doves; 89%), Ciconiiformes (storks; 78%), Coraciiformes (kingfishers and relatives; 76%), Cuculiformes (cuckoos and relatives; 73%), Pelicaniformes (pelicans and relatives; 71%), Falconiformes (diurnal birds of prey; 69%), and Passeriformes (perching birds; 59%) (Table S2). Strigiformes (owls) are spread between Groups 1 (54%) and 2 (46%), which is indicative of their varying degree of vulnerability depending on the species.
Vultures and hornbills, and other large avifauna such as eagles and bustards, are clearly the taxa most vulnerable to selective harvesting for the TM trade, and dominate the list of species with the highest total standardized variable scores (Table 9). Six of the nine vulture species recorded in TM markets across Africa are on our list of most vulnerable species (Table 9). These results confirm the observations of several studies that larger birds are more likely to be used for TM , ,  and/or more threatened by human persecution . Not only are they vulnerable to being utilised, but they also tend to have low population numbers and densities, and occupy more specific habitats – thus making them more likely to be naturally rare and constantly sparse in specific habitats within the harvesting catchments of hunters/harvesters/consumers.
Palearctic (PAL) Migrants to Africa
Palearctic migrant species are long-distance migrants that breed in Europe and North Africa but spend the boreal winter in sub-Saharan Africa , usually from October to April . Fifteen PAL migrants were recorded in the study, 14 between the markets of BF, BJ, NG and ZA (Table S1). Except for the European Roller (Coracias garrulous) and Barn Swallow (Hirundo rustica), 12 species were only recorded once and mainly in Benin. The only species to occur in South African markets was the large White Stork (Ciconia ciconia).
The species were mostly classified in Rabinowitz categories A–C, and had a mean rarity rank of 3.0 – indicating that they are generally not rare. The mean EOO for PALs at 13,537,786 km2 is twice that of the median range for all the species recorded in this study. However, the population trends for all except the Jack Snipe (Lymnocryptes minimus) are recorded as decreasing, with three species being classified as Near Threatened. Hence, the species are undergoing notable declines. As with the non-PAL species recorded in this study, birds with larger ranges tend to be smaller (mean = 389 g, or 170 g if the White Stork is excluded from the calculation).
Avifaunal Richness and Use in TM
Globally, there are about 10,064 bird species (extant and extinct) , , with 2,355–2,600 of these occurring in Africa, 145 of which are PAL migrants to Africa. Africa has 23% of the total global avifauna recognised as species by BirdLife International , , of which at least 354 species (17%) are used for TM across the continent. While there are only records for 306 species (13%) being sold in markets in seven countries, more species are utilised in more countries than have been reported in the literature. The authors have observed the remains of unidentified birds in the rural and urban markets of Mozambique, Malawi, South Africa, Swaziland and Zambia, and are aware of cross-border trade between several southern African countries, while a Nigerian study mentions observations of birds at markets in Guinea and the cross-border trade of vultures from Chad and Niger .
The paucity of information on bird utilisation in East Africa warrants further consideration. Is this a result of communities not using and trading avifauna in the same manner as communities in southern and West Africa, or is little ethno-ornithological research being conducted in the region? Our field observations suggest that East Africa is different with regards to bird species in trade. Their urban TM markets are tiny in proportion to city size compared to West and southern Africa, typically comprising small numbers of Maasai women vendors trading in a low diversity of plant species and with no bird species seen for sale. However, the lack of information does not necessarily infer that little trade in, or use of, birds occurs. An ornithologist currently working in Kenya indicated that Tanzania is the “epicentre for this [traditional healing] and no one works there, relatively speaking. Tanzanian witchdoctors are famous even in Kenya…It's pretty hard finding publications on this in East Africa” [Darcy Ogada, pers. comm. 14/02/2013]. Furthermore, Ogada & Kibuthu  report that the study of African owls, for example, is confounded by strong cultural prejudices associated with the use of owls for witchcraft, and that these collective prejudices have resulted in an almost complete lack of long-term studies, including in Kenya. What has emerged for the region is an escalating illicit trade in eggs of Mackinder's Eagle-owl (Bubo capensis mackinderi) from Kenya for ‘witchcraft’ and cancer, which are purportedly destined for Tanzania and the Middle East , . Clearly, consumptive utilisation of avifauna for TM in East Africa is occurring – but cultural taboos, a shortage of people conducting research, and differences in utilisation, trade and market structures/demand are factors limiting access to and availability of information. These same factors may also apply to other African countries for which there is also a paucity of information on consumptive use.
Despite this East African anomaly, there is a greater prevalence and diversity of bird species used and sold at TM markets in Africa than have been recorded on any other continent. In South America, for example, research on zootherapeutic resources in Brazil has indicated that there are at least 54 bird species used in folk medicinal practices , , ,  (which is one species more than the 53 species recorded in just one TM market in South Africa by ). The most specious orders used in Brazil that are also frequently found in African TM markets, are the Passeriformes, Galliformes, Columbiformes and Falconiformes (Table S4). Less specious in Brazil, but more common in African markets, are the Ciconiiformes and Piciformes. In Asia, there is a growing body of zootherapeutic studies being carried out among tribes across India that are reporting on the indigenous uses of animals for TM , –. Although the overall richness of bird species used in India is relatively low (at least 31 species recorded from 19 studies), avian taxa accounted for an average of 21% of the vertebrate species recorded per study. In common with Africa and South America, the most specious orders used in India are the Passeriformes and Galliformes (Table S4). However, unlike South America, there is also frequent utilisation of Strigiformes (owls), besides Bucerotiformes (hornbills) species that do not occur in the Americas, whereas much-used Tinamiformes (tinamous) only occur in the Americas.
The African continent is clearly a priority region for vulnerability assessments to take place since a larger proportion of the avifauna are utilised and sold for TM. At a national level, 26.6% (134) of Benin's 503 bird species, 23.6% (200) of the 848 bird species in Nigeria and 11.1% (84) of South Africa's 754 bird species are commercially traded for TM. By comparison, Brazil has 1,721 bird species and India 1,167 species , with use for TM respectively representing only 3.1% and 2.7% of the total number of bird species.
In contrast to India, no birds are recorded as commonly traded in Chinese TM , apart from one notable exception – the massive regional trade in swiftlet nests (Collocalia fuciphagus and C. maximus; Apodiformes) from Southeast Asia to China. These are marketed by such companies as Eu Yan Sang (www.euyansang.com) as a ‘health food’ and are sold in China as a luxury food (yàn wō), generally known as ‘bird's nest soup’. Trade in swiftlet nests are in the grey area between food and medicine , but this Asian trade is well studied and managed ,  and many of the nests come from artificially constructed nesting sites .
Risks to Frequently Traded Species in African Markets
The impact of the TM trade, and localised non-commercial medicinal consumption, on levels of avian vulnerability or resilience has rarely been quantitatively addressed. Cluster analysis and the Rabinowitz rarity model provided objective ways of assessing the vulnerability to selective harvesting for a large number of avifauna sold for TM by integrating biological and ecological traits with records of their trade.
Threatened species that are particularly targeted by TM hunters (e.g. vultures, ground-hornbills, nest-sealing hornbills and various eagles; Table 9) will experience more significant declines in population numbers if selective hunting persists. More recently, the illegal collection of Mackinder's Eagle-owl eggs for use in ‘witchcraft’ has become a major threat to the population in Kenya in , and locals were reportedly being paid in the range of USD300–500 per pair of eggs c.2011 [Darcy Ogada, pers. comm., 14/02/2013].
This study revealed that 24% of traded bird species are very common and locally abundant, in several habitats and over a large geographic area, and so are unlikely to be of conservation concern (‘A’, Table 6), while only 10% of species are in the category of most rare (H, Table 6). We also suggest that, compared to the multiple factors and mechanisms that threaten avian species (e.g. intrinsic biological features, habitat loss, deleterious farming practices, and other forms of utilisation) , , , , the TM trade plays a minor role in the decline of species. Even in the case of birds in families prone to extinction through human use that are also used for TM, such as penguins (Spheniscidae), and albatrosses and petrels (Procellariidae) , our impressions are that these birds probably died at sea before ending up on beaches. For example, the presence of such species as the Shy Albatross (Thalassarche cauta) in a South African market (Table S1) may be the indirect result of natural or unnatural events, such as bad weather or long-line casualties, rather than targeted harvesting events. Based on discussions with traditional healers, the use of these stranded seabirds is influenced by the ritual potency of what are, to the healers, unusual pelagic birds tossed out by the sea and an uncommon occurrence.
In their study of the ecological basis for extinction risks in birds, Owens & Bennett  made the important point that different bird lineages may follow different paths to extinction. One route is for large-bodied, slow-breeding bird species to become threatened when the fecundity-mortality balance is disrupted by, for example, human use or introduced predators. A second route is for ecologically specialized species to become threatened by habitat loss . These risk factors also apply to birds used for African TM. But whereas Owens & Bennet  correlated extinction risks incurred by smaller birds with a high degree of habitat specialisation, we show in Figure 5 and Table 8 that larger birds used for TM are naturally rarer and occur in narrower and more specialised habitats (category H). Furthermore, the most common birds are generally smaller and occur in a broader range of habitats (category A) (particularly if ostriches are excluded as an outlier). And, while penguins are rarely used and probably opportunistically acquired in Africa, the specific trade in many Accipitridae, owls, large hornbills (particularly Abyssinian and Southern Ground-hornbills) and, in such countries as Cote d'Ivoire, heads of Black-crowned Crane (Balearica pavonina), is cause for concern (Tables 3 & 9). Furthermore, vulnerable nest types and sites, combined with small clutch sizes may be a compounding factor in the vulnerability of species to selective harvesting.
It is widely accepted, for example, that birds using closed nests (such as natural or excavated cavities in trees) are prone to lower levels of natural nest predation . However, adults using closed nests may be more easily trapped/killed as the sites are often re-used and accordingly the species are more vulnerable to detection and predation by people. Nearly half of the 19 most-frequently traded bird species recorded in African TM markets (hoopoes, hornbills, parrots, rollers and owls), nest in tree cavities (Table 4). While human predation on hole- and open-nesting bird species is a conservation challenge, the behaviour of some species also offers an opportunity for harvest and rearing for TM of redundant second-hatched chicks normally starved or killed by the first-hatched elder sibling , and provision of safe artificial nests 
In addition to threats from the African TM trade, many families are under selective pressure from anthropogenic factors such as collisions with aircraft and overhead power lines and, most importantly, poisoning. The Accipitridae, especially vultures, are frequently victims of these threats , , and large-scale mortalities have been caused by deliberate and inadvertent poisoning with such pesticides as the carbamate-based pesticide Furadan –. In South Africa, poisoning has been the main factor causing a decline in Bateleur Eagle  and maybe Tawny Eagle (Aquila rapax) and Southern Ground-hornbill populations . In Kenya, pesticides have been implicated in the decline of several species and populations, including raptors and Mackinder's Eagle-owls , , .
Effective and appropriate conservation strategies for utilised species should ideally be based on a thorough understanding of the world views of the users, such as why particular birds are selected across diverse African healing traditions, together with a thorough understanding of the species' biology and the TM supply chains for those species. An appreciation of the cultural and religious context is essential to develop viable conservation strategies. Selection of particular groups of birds is based on similar traits or combinations of characteristics that have symbolic meaning. Examples of these factors are the large body size and red, black and white colours of both ground-hornbill species and Bateleur eagles. In Nigeria, the raucous calls of Western Grey Plantain-eaters (Crinifer piscator) and three turaco species (Violet, Guinea and Yellow-billed) (Musophaga violacea, Tauraco persa, T. macrorhynchus) are believed to attract customers, the melodious calls of the Black-crowned Tchagra (Tchagra senegalensis) to impart musical ability . It is well known from distributions of bird species that African avifauna vary from country to country, so it is not surprising that different species within the same groups of birds sharing common characteristics are used. Across Africa, however, birds are ritually potent symbols that have metaphorical meaning in healing and religious traditions that link mind and body, and this potency connects to widespread beliefs in the power of wild places and the meaning of natural events. In the Congo, for example, 46 bird species were documented that Mbuti hunter-gatherers believe are mediators between the spirit world and human society, and 20 bird species with the power to cause illness . It is a mistake to think that Mbuti beliefs are archaic and isolated, since our work shows that belief in the ritual potency of birds is widespread. It is not by chance that the Great Blue Turacos (Corythaeola cristata), Senegal Coucals (Centropus senegalensis) and egrets, that the Mbuti consider dangerously powerful, are also sold in urban markets. The eerie calls of nocturnal birds such as owls are a further example, with widespread symbolic beliefs attributed to a range of owl species , , . The similarity of use of avifaunal orders is high, especially at a morphospecies level (Table 5), suggesting an opportunity for a common understanding of what drives demand across the continent. This also suggests common links across African belief systems, and offers an opportunity for similar conservation and resource-management strategies across a wide range of countries that recognise cultural links to African bird diversity.
Bird conservation policies and practices in Africa need to take into account bird use for traditional purposes, particularly in national and regional conservation strategies across West and southern Africa. The selection of taxa for TM involves layers of anthropological, ecological, behavioural and phenotypic complexity. Across Africa, large and/or conspicuous birds are targeted by the TM trade. Since bigger birds tend to occur at lower population densities and have slower reproductive rates, which makes them rarer and with lower population numbers per unit area of suitable habitat, their populations are more vulnerable, especially if their habitats are also being transformed. Vultures and hornbills, and such other large avifauna as eagles and bustards, are the most vulnerable taxa. Selective harvesting of these vulnerable birds for the TM trade can no longer be ignored. Bird conservation strategies need to take the TM trade into account, starting with the 19 species listed in this study (Table 9).
It must be noted that the TM trade in wild birds is only a part, probably a relatively small one, of the commercial trade in wild birds in Africa. Other important forms of trade include birds hunted primarily for food ('bush meat') and, probably most voluminous of all, trade in live birds for aviculture and pets. Each form of trade is probably to an extent integrated and interdependent, since all involve initially hunters and finally traders. For example, food items might end up in TM if not consumed while fresh, and pre-export casualties from the live-bird trade are potentially available as food while fresh and later for TM. The different forms and proportions of each trade type probably vary by bird taxa and country, for example with bush meat trade expected for larger birds in forested areas and live-bird trade best known for parrots and finches and from Tanzania and Senegal. It would be revealing, therefore, to have similar analyses of use by species and country conducted for these non-TM forms of trade, so that a comprehensive understanding can be established of how these different trades integrate into their combined effect on avifaunal conservation.
The mean mass (g) of species sold in ZA, NG and BJ markets, and the total for seven countries combined (‘All traded’), in each Rabinowitz category of commonness and rarity (excluding ostrich). Birds from ZA tended to be larger and were almost twice as heavy as birds from BJ and NG.
Inventory of avian species used and sold for traditional medicine in 25 African countries with their IUCN conservation statuses, Rabinowitz rarity classification, population trend, and the number of countries in which they were recorded to be sold in markets. All abbreviations are listed as footnotes at the end of the table. The taxonomic classification follows BirdLife.
a,b,c. Conservation priority groupings, and hence vulnerability to exploitation, of species sold in traditional medicine markets across Africa. Species were sorted into two groups of higher and lower priority and vulnerability according to the K-means cluster analysis. Species are listed within the groups in descending order of their maximum scores for the three standardised variables.
Mean EOOs for traded bird species in the eight Rabinowitz and four rarity categories (corresponding with the classes in Table 1), where A and 4 are the most common species, and H and 1 are most rare species (excludes PALs and exotics). Birds with an EOO >6,790,00 km2 were assigned to classes A to D.
We acknowledge Peter Howard for use of his photographs taken in Morocco and Togo. Kaera Coetzer is thanked for making the map.
Conceived and designed the experiments: VLW ABC. Analyzed the data: VLW. Contributed to the writing of the manuscript: VLW ABC ACK RKB.
- 1. Graham-Rowe D (2011) Biodiversity: endangered and in demand. Nature 480: S101–S103 doi: 10.1038/480S101a.
- 2. Costa-Neto EM (1999) Healing with animals in Feira de Santana city, Bahia, Brazil. Journal of Ethnopharmacology 65: 225–230 doi: 10.1016/S0378-8741(98)00158-5.
- 3. Costa-Neto EM (2005) Animal-based medicines: biological prospection and the sustainable use of zootherapeutic resources. Anais da Academia Brasileira de Ciências 77: 33–43 doi: 10.1590/S0001-37652005000100004.
- 4. Lev E (2003) Traditional healing with animals (zootherapy): medieval to present-day Levantine practice. Journal of Ethnopharmacology 85: 107–118 doi: 10.1016/S0378-8741(02)00377-X.
- 5. Butchart SHM (2008) Red List Indices to measure the sustainability of species use and impacts of invasive alien species. Bird Conservation International 18: S245–S262 doi: 10.1017/S095927090800035X.
- 6. France H (1905) Customs of the Awuna tribes. Journal of the Royal African Society 5(17): 38–40 Available: http://www.jstor.org/stable/715151.
- 7. Garbutt HW (1909) Native witchcraft and superstition in South Africa. The Journal of the Royal Anthropological Institute of Great Britain and Ireland 39: 530–558 doi: 10.2307/2843218.
- 8. Hollis C (1901) Notes on the history and customs of the people of Taveta, East Africa. Journal of the Royal African Society 1(1): 98–125 Available: http://www.jstor.org/stable/714706 Accessed 2014 Aug 3..
- 9. Alves RRN, Rosa IL, Santana GG (2007) The role of animal-derived remedies as complementary medicine in Brazil. BioScience 57: 949–955 doi: 10.1641/B571107.
- 10. Alves RRN, Rosa IL (2010) Trade of animals used in Brazilian traditional medicine: trends and implications for conservation. Human Ecology 38: 691–704 doi: 10.1007/s10745-010-9352-0.
- 11. Mahawar MM, Jaroli DP (2008) Traditional zootherapeutic studies in India: a review. Journal of Ethnobiology and Ethnomedicine 4: 17 doi: 10.1186/1746-4269-4-17.
- 12. Anyinam C (1995) Ecology and ethnomedicine: exploring links between current environmental crisis and indigenous medical practices. Social Science and Medicine 40: 321–329 doi: 10.1016/0277-9536(94)E0098-D.
- 13. Cocker M, Mikkola H (2001) Magic, myth and misunderstanding: cultural responses to owls in Africa and their implications for conservation. Bulletin of the African Bird Club 8(1): 30–35.
- 14. Bruyns RK, Williams VL, Cunningham AB (2013) Finely ground-hornbill: the sale of Bucorvus cafer in a traditional medicine market in Bulawayo, Zimbabwe. In: Alves RRN, Rosa IL, editors. Animals in Traditional Folk Medicine. Berlin Heidelberg: Springer-Verlag. pp. 475–486.
- 15. Adjakpa JB, Tchabi A, Ogouvide FT (2002) Oiseaux utilisés en pharmacopée traditionnelle au Bénin. Malimbus 24: 1–14 Available: http://malimbus.free.fr/articles/V24/24001014.pdf Accessed 2014 Aug 3..
- 16. Adeola MO (1992) Importance of wild animals and their parts in the culture, religious festivals, and traditional medicine, of Nigeria. Environmental Conservation 19: 125–134 doi: 10.1017/S0376892900030605.
- 17. Nikolaus G (2001) Bird exploitation for traditional medicine in Nigeria. Malimbus 23: 45–55 Available: http://malimbus.free.fr/articles/V23/23045055.pdf Accessed 2014 Aug 3..
- 18. Cunningham AB, Zondi AS (1991) Use of Animal Parts for the Commercial Trade in Traditional Medicines. Unpublished Report. Pietermaritzburg: Institute of Natural Resources.
- 19. Mander M, Diederichs N, Ntuli L, Mavundla K, Williams V, et al.. (2007) Survey of the Trade in Vultures for the Traditional Health Industry in South Africa. Unpublished Report. Durban: FutureWorks.
- 20. Whiting MJ, Williams VL, Hibbitts TJ (2011) Animals traded for traditional medicine at the Faraday market in South Africa: species diversity and conservation implications. Journal of Zoology 284: 84–96 doi: 10.1111/j.1469-7998.2010.00784.x.
- 21. Marshall NT (1998) Searching for a Cure: Conservation of Medicinal Wildlife Resources in East and Southern Africa. Cambridge, Traffic International.
- 22. Williams VL, Cunningham AB, Bruyns RK, Kemp AC (2013) Birds of a feather: quantitative assessments of the diversity and levels of threat to birds used in African traditional medicine. In: Alves RRN, Rosa IL, editors. Animals in Traditional Folk Medicine. Berlin Heidelberg: Springer-Verlag. pp. 383–420.
- 23. Baskind R, Birbeck G (2005) Epilepsy care in Zambia: a study of traditional healers. Epilepsia 46: 1121–1126 doi: 10.1111/j.1528-1167.2005.03505.x.
- 24. Beninguisse G, De Brouwere V (2004) Tradition and modernity in Cameroon: the confrontation between social demand and biomedical logistics of health services. African Journal of Reproductive Health 8: 152–175 doi: 10.2307/3583401.
- 25. Buckley AD (1985) Yoruba Medicine. New York: Oxford University Press.
- 26. Cocker M (2000) African birds in traditional magico-medicinal use – a preliminary survey. Bulletin of the African Bird Club 7(1): 60–65.
- 27. Crump CM (2003) Need to assess animals used by traditional healers in South Africa. In: Hawkers of Health: an Investigation of the Faraday Street Traditional Medicine Market in Johannesburg. Appendix 4. Williams VL, Unpublished Report to Gauteng Directorate for Nature Conservation, Johannesburg. pp 44–46. Available: http://wiredspace.wits.ac.za/bitstream/handle/10539/5313/Faraday%20Report%20Appendix.pdf?sequence=16.
- 28. Cunningham AB (1990) Vultures and the trade in traditional medicine. Vulture News 24: 3–10.
- 29. Derwent S, Mander M (1997) Twitchers bewitched. The use of birds in traditional healing. Africa: Birds & Birding 2(1): 22–25.
- 30. El-Kamali HH (2000) Folk medicinal use of some animal products in Central Sudan. Journal of Ethnopharmacology 72: 279–282 doi: 10.1016/S0378-8741(00)00209-9.
- 31. Nevadomsky J (1988) Kemwin-Kemwin: the apothecary shop in Benin City. African Arts 22: : 72–83,100. doi: 10.2307/3336693.
- 32. Haileselasie TH (2012) Traditional zootherapeutic studies in Degu'a Tembien, Northern Ethiopia. Current Research Journal of Biological Sciences 4: 563–569.
- 33. Kizungu B, Ntabaza M, Mburunge M (1998) Ethno-ornithology of the Tembo in the eastern DRC (former Zaire): Part one, Kalehe Zone. African Study Monographs 19: 102–113.
- 34. Lacaille J (1951) Mai Mai, Harley Street of Jo'burg's "muti" men. Outspan 9: : 47–49, 81.
- 35. Loeb EM, Koch C, Loeb E-MK (1956) Kuanyama Ambo magic. 6. Medicinal, cosmetical, and charm flora and fauna. The Journal of American Folklore 69: 147–174 doi: 10.2307/537275.
- 36. Macdonald IAW (1985) Witch-doctors versus wildlife in southern Africa. African Wildlife 38(1): 4–9.
- 37. Ngwenya MP (2001) Implications for the Medicinal Animal Trade for Nature Conservation in KwaZulu-Natal. Unpublished Ezimvelo KZN Wildlife Report No. NA/124/04.
- 38. Ntiamoa-Baidu Y (1987) West African wildlife: a resource in jeopardy. Unasylva (FAO) 39: 27–35.
- 39. Ogada DL (2007) The Ecology and Conservation of Mackinder's Eagle Owls (Bubo capensis mackinderi) in Central Kenya in Relation to Agricultural Land-use and Cultural Attitudes. Thesis, Grahamstown: Rhodes University. Available: http://eprints.ru.ac.za/1290/. Accessed 2014 Aug 3.
- 40. Serge BK (2010) Birds, the Banso people of northwestern Cameroon and their culture. Abstracts from the 12th ISE – Recent Developments in Ethnoornithology. Torfino, Canada. May 2010. Available: http://uk.groups.yahoo.com/group/Ethnoornithology/message/579. Accessed 2014 Aug 3.
- 41. Serle W, Morel GJ (1979) Les Oiseaux de l'Ouest Africain. Neuchâtel: Delachaux et Niestlé. 331 p.
- 42. Simelane TS, Kerley GIH (1998) Conservation implications of the use of vertebrates by Xhosa traditional healers in South Africa. South African Journal of Wildlife Research 28: 121–126.
- 43. Sodeinde OA, Soewu DA (1999) Pilot study of the traditional medicine trade in Nigeria. TRAFFIC Bulletin 18: 35–40.
- 44. Soewu DA (2008) Wild animals in ethnozoological practices among the Yorubas of southwestern Nigeria and the implications for biodiversity conservation. African Journal of Agricultural Research 3: 421–427 Available: http://www.academicjournals.org/article/article1380788373_soewu.pdf Accessed 2014 Aug 3..
- 45. Taylor ME, Fox J (1992) The fetish market, Lomé, Togo. The Nigerian Field 57: 119–125.
- 46. White RM, Cocks M, Herbert DG, Hamer ML (2004) Traditional medicines from forest animals. In: Lawes MJ, Eeley HAC, Shackleton CM, Geach BGS, editors. Indigenous Forests and Woodlands in South Africa: People, Policy and Practice. Pietermaritzburg: University of Natal Press. pp. 474–477.
- 47. Tobias JA, Seddon N, Spottiswoode CN, Pilgrim JD, Fishpool LDC, et al. (2010) Quantitative criteria for species delimitation. Ibis 152: 724–746 doi: 10.1111/j.1474-919X.2010.01051.x.
- 48. Del Hoyo J, Collar NJ, Christie DA, Elliot A, Fishpool LDC (2014) Illustrated Checklist of the Birds of the World. Volume 1: Non-passerines. HBW and BirdLife International. In press.
- 49. BirdLife International (2008) Palearctic-African migratory birds have suffered substantial declines. Presented as part of the BirdLife State of the world's birds website. http://birdlife.org/datazone/sowb/casestudy/66. Accessed 2014 Aug 3.
- 50. Vickery JA, Ewing SR, Smith KW, Pain DJ, Bairlein F, et al. (2014) The decline of Afro-Palaearctic migrants and an assessment of potential causes. Ibis 156: 1–22 doi: 10.1111/ibi.12118.
- 51. Colwell R (2006) EstimateS: statistical estimation of species richness and shared species from samples, version 7.5.1. Available: http://viceroy.eeb.uconn.edu/estimates. Accessed 2009 July 1.
- 52. Rabinowitz D (1981) Seven forms of rarity. In: Synge H, editor. The Biological Aspects of Rare Plant Conservation. Chichester: John Wiley & Sons. pp. 205–217.
- 53. Rabinowitz D, Cairns S, Dillon T (1986) Seven forms of rarity and their frequency in flora of the British Isles. In: Soule ME, editor. Conservation Biology, the Science and Scarcity of Diversity. Sunderland, Massachusetts: Sinauer. pp. 182–204.
- 54. Yu J, Dobson FS (2000) Seven forms of rarity in mammals. Journal of Biogeography 27: 131–139 doi: 10.1046/j.1365-2699.2000.00366.x.
- 55. Sinclair I, Ryan P (2010) Chamberlain's Birds of Africa South of the Sahara. Cape Town: Struik.
- 56. Owens IPF, Bennett PM (2000) Ecological basis of extinction risk in birds: habitat loss versus human persecution and introduced predators. Proceedings of the National Academy of Sciences 97: 12144–12148 doi: 10.1073/pnas.200223397.
- 57. Chittenden H, Upfold G (2007) Roberts Bird Guide. Cape Town: Jacana.
- 58. Peters RH (1983) The Ecological Implications of Body Size. Cambridge: Cambridge University Press. Available: http://dx.doi.org/10.1017/CBO9780511608551.
- 59. Given DR, Norton DA (1993) A multivariate approach to assessing threat and for priority setting to threatened species conservation. Biological Conservation 64: 57–66 doi: 10.1016/0006-3207(93)90383-C.
- 60. Hall AV (1993) Setting conservation priorities for threatened species: a joint grouping and sequencing method. South African Journal of Botany 59: 581–591. doi: 10.1016/0006-3207(95)93792-b
- 61. Ding CG, Woo YY, Sheu H-J, Chien H-C, Shen S-F (1996) An effective statistical approach for comparative risk assessment. Risk Analysis 16: 411–419 doi: 10.1111/j.1539-6924.1996.tb01475.x.
- 62. Myers OB (1999) On aggregating species for risk assessment. Human and Ecological Risk Assessment: An International Journal 5: 559–574 doi: 10.1080/10807039991289536.
- 63. Simeonov V (2003) Environmetric strategies to classify, interpret and model risk assessment and quality of environmental systems. Clean Technologies and Environmental Policy 5: 190–199 doi: 10.1007/s10098-003-0201-3.
- 64. Williams VL (2007) The Design of a Risk Assessment Model to Determine the Impact of the Herbal Medicine Trade on the Witwatersrand on Resources of Indigenous Plant Species. PhD thesis, University of the Witwatersrand, Johannesburg. http://hdl.handle.net/10539/5313. Accessed 2014 Aug 3.
- 65. Ramos MC (2001) Divisive and hierarchical clustering techniques to analyse variability of rainfall distribution patterns in a Mediterranean region. Atmospheric Research 57: 123–138 doi: 10.1016/S0169-8095(01)00065-5.
- 66. Höft M, Barik SK, Lykke AM (1999) Quantitative Ethnobotany. Applications of Multivariate and Statistical Analyses in Ethnobotany. People and Plants Working Paper 6. UNESCO, Paris. Available: http://pure.au.dk/portal/files/17477574/hoeftbariklykke1999.pdf. Accessed 2014 Aug 3.
- 67. BirdLife International (2012) Barn Owl. Available: http://www.birdlife.org/. Accessed 2012 Dec 2.
- 68. Julliard R, Jiguet F, Couvet D (2003) Common birds facing global changes: what makes a species at risk? Global Change Biology 10: 148–154 doi: 10.1111/j.1365-2486.2003.00723.x.
- 69. BirdLife International (2008) The status of the world's birds has deteriorated over the last 20 years. Presented as part of the BirdLife State of the world's birds website. http://birdlife.org/datazone/sowb/casestudy/72. Accessed 2014 Aug 3.
- 70. BirdLife International (2013) State of Africa's birds: outlook for our changing environment. BirdLife International Africa Partnership, Nairobi, Kenya. http://www.birdlife.org/datazone/userfiles/file/sowb/pubs/State_of_Africas_Birds_report_2013_(FINAL).pdf. Accessed 2014 Aug 3.
- 71. BirdLife International (2014) Taxonomy. Available: http://www.birdlife.org/datazone/info/taxonomy. Accessed 2014 Jun 25.
- 72. Ogada DL, Kibuthu PM (2012) Breeding ecology of Mackinder's Eagle-Owls (Bubo capensis mackinderi) in farmlands of central Kenya. Journal of Raptor Research 46: 327–335 doi: 10.3356/JRR-11-90.1.
- 73. Downs C, Harebottle D, Dodman T, Ndiaye A, Barnard P, et al. (2014) Birds in a changing environment: Report on the 13th Pan-African Ornithological Congress in Arusha, Tanzania. Ostrich 85: iii–vi doi: 10.2989/00306525.2014.912452.
- 74. The Standard (2011) Witches spell doom to bird of omen. Available: http://www.standardmedia.co.ke/?articleID=2000039250&story_title=Witches-spell-doom-to-bird-of-omen. Accessed 22 June 2014.
- 75. Bezerra DMM, de Araujo HFP, Alves AGC, Alves RRN (2013) Birds and people in semiarid northeastern Brazil: symbolic and medicinal relationships. Journal of Ethnobiology and Ethnomedicine 9: 3 doi: 10.1186/1746-4269-9-3.
- 76. Costa-Neto EM, Alves RRN (2010) Estado da arte da zooterapia popular no Brasil. In: Costa-Neto EM, Alves RRN, editors. Zooterapia: os Animais na Medicine Populair Brasiliera. Recife: Nupeea. pp. 15–54.
- 77. Betlu ALS (2013) Indigenous knowledge of zootherapeutic use among the Biate tribe of Dima Hasao District, Assam, Northeastern India. Journal of Ethnobiology and Ethnomedicine 9: 56 doi: 10.1186/1746-4269-9-56.
- 78. Chakravorty J, Meyer-Rochow VB, Ghosh S (2011) Vertebrates used for medicinal purposes by members of the Nyishi and Galo tribes in Arunachal Pradesh (North-East India). Journal of Ethnobiology and Ethnomedicine 7: 13 doi: 10.1186/1746-4269-7-13.
- 79. Jaroli DP, Mahawar MM, Vyas N (2010) An ethnozoological study in the adjoining areas of Mount Abu wildlife sanctuary, India. Journal of Ethnobiology and Ethnomedicine 6: 6 doi: 10.1186/1746-4269-6-6.
- 80. Mishra N, Rout SD, Panda T (2013) Ethno-zoological studies and medicinal values of Similipal Biosphere Reserve, Orissa, India. African Journal of Pharmacy and Pharmacology 5: 6–11 doi: 10.5897/AJPP09.241.
- 81. BirdLife International (2014) Country profiles. Available: http://www.birdlife.org/datazone/country. 2014 Feb 24.
- 82. Zhao Z (2004) An Illustrated Chinese Materia Medica in Hong Kong. Hong Kong: Chun Hwa Book Company.
- 83. Etkin NL, Ross PJ (1982) Food as medicine and medicine as food: an adaptive framework for the interpretation of plant utilization among the Hausa of northern Nigeria. Social Science and Medicine 16: 1559–1573 doi: 10.1016/0277-9536(82)90167-8.
- 84. Hobbs JJ (2003) Problems in the harvest of edible birds' nests in Sarawak and Sabah, Malaysian Borneo. Biodiversity Conservation 13: 2209–2226 doi: 10.1023/B:BIOC.0000047905.79709.7f.
- 85. Sankaran R (2001) The status and conservation of the Edible-nest Swiftlet (Collocalia fuciphaga) in the Andaman and Nicobar Islands. Biological Conservation 97: 283–294 doi: 10.1016/S0006-3207(00)00124-5.
- 86. Saengkrajang W, Matan N, Matan N (2013) Nutritional composition of the farmed edible bird's nest (Collocalia fuciphaga) in Thailand. Journal of Food Composition and Analysis 31: 41–45 doi: 10.1016/j.jfca.2013.05.001.
- 87. Machado N, Loyola RD (2013) A comprehensive quantitative assessment of bird extinction risk in Brazil. PLoS ONE 8: e72283 doi: 10.1371/journal.pone.0072283.
- 88. Ogada DL, Kibuthi PM (2008) Conserving Mackinder's eagle owls in farmlands of Kenya: assessing the influence of pesticide use, tourism and local knowledge of owl habits in protecting a culturally loathed species. Environmental Conservation 35: 252–260 doi: 10.1017/S0376892908005006.
- 89. Martin TE, Li P (1992) Life history traits of open- vs. cavity-nesting birds. Ecology 73: 579–592 doi: 10.2307/1940764.
- 90. Kemp A (2000) The sustainable utilisation of birds. Emu 100: 355–365 doi: 10.1071/MU0003S.
- 91. Wilson G, Hockey PAR (2013) Causes of variable reproductive performance by Southern Ground-hornbill Bucorvus leadbeateri and implications for management. Ibis 155: 476–484 doi: 1111/ibi.12042.
- 92. Herholdt JJ, Anderson MD (2006) Observations on the population and breeding status of the African White-backed Vulture, the Black-chested Snake Eagle, and the Secretarybird in Kgalagadi Transfrontier Park. Ostrich: Journal of African Ornithology 77: 127–135 doi: 10.2989/00306520609485523.
- 93. Otieno PO, Lalah JO, Virani M, Jondiko IO, Schramm K-W (2010) Carbofuran and its toxic metabolites provide forensic evidence for Furadan exposure in vultures (Gyps africanus) in Kenya. Bulletin of Environmental Contamination and Toxicology 84: 536–544 doi: 10.1007/s00128-010-9956-5.
- 94. Ogada DL, Keesing F (2010) Decline of raptors over a three-year period in Laikipia, central Kenya. Journal of Raptor Research 44: 129–135 doi: 10.3356/JRR-09-49.1.
- 95. de Kock AC, Watson RT (1985) Organochlorine residue levels in Bateleur eggs from the Transvaal. Ostrich 56: 278–280.
- 96. Barnes KN (2000) The Eskom Red Data Book of Birds of South Africa, Lesotho and Swaziland. Johannesburg: BirdLife South Africa.
- 97. Ichikawa M (1998) The birds as indicators of the invisible world: ethno-ornithology of the Mbuti hunter-gatherers. African Study Monographs 25 (Suppl): 105–121. Available: http://hdl.handle.net/2433/68390. Accessed 2014 Aug 3.
- 98. Marcot BG (2007) Owls in native cultures of central Africa and North America. Tyto Newsbrief 11: 5–9.
- 99. Cunningham AB (2001) Applied Ethnobotany: People, Wild Plant Use and Conservation. London: Earthscan.